osmo-msc/src/libmsc/gsm_04_08.c

1936 lines
55 KiB
C
Raw Normal View History

/* GSM Mobile Radio Interface Layer 3 messages on the A-bis interface
* 3GPP TS 04.08 version 7.21.0 Release 1998 / ETSI TS 100 940 V7.21.0 */
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
/* (C) 2008-2016 by Harald Welte <laforge@gnumonks.org>
* (C) 2008-2012 by Holger Hans Peter Freyther <zecke@selfish.org>
*
* All Rights Reserved
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License as published by
* the Free Software Foundation; either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Affero General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*
*/
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <stdbool.h>
#include <errno.h>
#include <time.h>
2008-12-27 01:55:51 +00:00
#include <netinet/in.h>
#include <regex.h>
#include <sys/types.h>
#include "bscconfig.h"
#include <osmocom/msc/db.h>
#include <osmocom/msc/debug.h>
#include <osmocom/msc/gsm_data.h>
#include <osmocom/msc/gsm_subscriber.h>
#include <osmocom/msc/gsm_04_11.h>
#include <osmocom/msc/gsm_04_08.h>
#include <osmocom/msc/gsm_04_80.h>
#include <osmocom/msc/gsm_04_14.h>
#include <osmocom/msc/gsm_09_11.h>
#include <osmocom/msc/signal.h>
#include <osmocom/msc/transaction.h>
#include <osmocom/msc/silent_call.h>
#include <osmocom/msc/mncc_int.h>
src: port openBSC over libosmo-abis This is a big patch that ports openBSC over libosmo-abis. Sorry, the changes that are included here are all dependent of libosmo-abis, splitting them into smaller pieces would leave the repository in some intermediate state, which is not desired. The main changes are: - The directory libabis/ has been removed as it now lives in libosmo-abis. - new configuration file format for nanoBTS and HSL femto, we need to define the virtual e1_line and attach it to the OML link. - all the existing BTS drivers (nanoBTS, hsl femto, Nokia site, BS11 and rbs2000) now use the new libosmo-abis framework. - use r232 input driver available in libosmo-abis for bs11_config. - use ipa_msg_recv instead of old ipaccess_read_msg function. - delete definition of gsm_e1_subslot and input_signal_data. These structures now lives in libosmo-abis. Most of this patch are deletions of libabis/ which has been moved to libosmo-abis. This patch also modifies openBSC to use all the new definitions available in libosmocore and libosmo-abis. In order to do that, we have replaced the following: - DINP, DMI, DMIB and DMUX by their respective DL* correspondences. - SS_GLOBAL by SS_L_GLOBAL - SS_INPUT by SS_L_INPUT - S_GLOBAL_SHUTDOWN by S_L_GLOBAL_SHUTDOWN - SS_INPUT by SS_L_INPUT - S_INP_* by S_L_INP_* sub-signals - E1INP_NODE by L_E1INP_NODE vty node This patch has been tested with: - one nanoBTS - the HSL femto with the examples available under libosmo-abis - BS11 with both dahdi and misdn drivers.
2011-08-17 20:44:07 +00:00
#include <osmocom/abis/e1_input.h>
#include <osmocom/core/bitvec.h>
#include <osmocom/msc/vlr.h>
#include <osmocom/msc/msc_ifaces.h>
#include <osmocom/gsm/gsm48.h>
#include <osmocom/gsm/gsm0480.h>
#include <osmocom/gsm/gsm_utils.h>
#include <osmocom/gsm/protocol/gsm_04_08.h>
#include <osmocom/core/msgb.h>
#include <osmocom/core/talloc.h>
#include <osmocom/core/utils.h>
#include <osmocom/core/byteswap.h>
#include <osmocom/gsm/tlv.h>
#include <osmocom/crypt/auth.h>
#ifdef BUILD_IU
#include <osmocom/ranap/iu_client.h>
#endif
#include <osmocom/msc/msc_ifaces.h>
#include <osmocom/msc/a_iface.h>
#include <osmocom/msc/msc_mgcp.h>
#include <assert.h>
mscsplit: various preparations to separate MSC from BSC Disable large parts of the code that depend on BSC presence. The code sections disabled by #if BEFORE_MSCSPLIT shall be modified or dropped in the course of adding the A-interface. Don't set msg->lchan nor msg->dst. Don't use lchan in libmsc. Decouple lac from bts. Prepare entry/exit point for MSC -> BSC and MSC -> RNC communication: Add msc_ifaces.[hc], a_iface.c, with a general msc_tx_dtap() to redirect to different interfaces depending on the actual subscriber connection. While iu_tx() is going to be functional fairly soon, the a_tx() is going to be just a dummy for some time (see comment). Add Iu specific fields in gsm_subscriber_connection: the UE connection pointer and an indicator for the Integrity Protection status on Iu (to be fully implemented in later commits). Add lac member to gsm_subscriber_connection, to allow decoupling from bts->location_area_code. The conn->lac will actually be set in iu.c in an upcoming commit ("add iucs.[hc]"). move to libcommon-cs: gsm48_extract_mi(), gsm48_paging_extract_mi(). libmsc: duplicate gsm0808 / gsm48 functions (towards BSC). In osmo-nitb, libmsc would directly call the functions on the BSC level, not always via the bsc_api. When separating libmsc from libbsc, some functions are missing from the linkage. Hence duplicate these functions to libmsc, add an msc_ prefix for clarity, also add a _tx to gsm0808_cipher_mode(): * add msc_gsm0808_tx_cipher_mode() (dummy/stub) * add msc_gsm48_tx_mm_serv_ack() * add msc_gsm48_tx_mm_serv_rej() Call these from libmsc instead of * gsm0808_cipher_mode() * gsm48_tx_mm_serv_ack() * gsm48_tx_mm_serv_rej() Also add a comment related to msc_gsm0808_tx_cipher_mode() in two places. Remove internal RTP streaming code; OsmoNITB supported that, but for OsmoMSC, this will be done with an external MGCP gateway. Remove LCHAN_MODIFY from internal MNCC state machine. Temporarily disable all paging to be able to link libmsc without libbsc. Skip the paging part of channel_test because the paging is now disabled. Employ fake paging shims in order for msc_vlr_tests to still work. msc_compl_l3(): publish in .h, tweak return value. Use new libmsc enum values for return val, to avoid dependency on libbsc headers. Make callable from other scopes: publish in osmo_msc.h and remove 'static' in osmo_msc.c add gsm_encr to subscr_conn move subscr_request to gsm_subscriber.h subscr_request_channel() -> subscr_request_conn() move to libmsc: osmo_stats_vty_add_cmds() gsm_04_08: remove apply_codec_restrictions() gsm0408_test: use NULL for root ctx move to libbsc: gsm_bts_neighbor() move to libbsc: lchan_next_meas_rep() move vty config for t3212 to network level (periodic lu) remove unneccessary linking from some tests remove handle_abisip_signal() abis_rsl.c: don't use libvlr from libbsc gsm_subscriber_connection: put the LAC here, so that it is available without accessing conn->bts. In bsc_api.c, place this lac in conn for the sake of transition: Iu and A will use this new field to pass the LAC around, but in a completely separate OsmoBSC this is not actually needed. It can be removed again from osmo-bsc.git when the time has come. Siemens MRPCI: completely drop sending the MRPCI messages for now, they shall be added in osmo-bsc once the A-Interface code has settled. See OS#2389. Related: OS#1845 OS#2257 OS#2389 Change-Id: Id3705236350d5f69e447046b0a764bbabc3d493c
2017-05-08 13:12:20 +00:00
void *tall_locop_ctx;
void *tall_authciphop_ctx;
2009-06-20 20:36:41 +00:00
static int gsm0408_loc_upd_acc(struct ran_conn *conn,
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
uint32_t send_tmsi);
/*! Send a simple GSM 04.08 message without any payload
* \param conn Active RAN connection
* \param[in] pdisc Protocol discriminator
* \param[in] msg_type Message type
* \return result of \ref gsm48_conn_sendmsg
*/
int gsm48_tx_simple(struct ran_conn *conn,
uint8_t pdisc, uint8_t msg_type)
{
struct msgb *msg = gsm48_msgb_alloc_name("GSM 04.08 TX SIMPLE");
struct gsm48_hdr *gh = (struct gsm48_hdr *) msgb_put(msg, sizeof(*gh));
gh->proto_discr = pdisc;
gh->msg_type = msg_type;
return gsm48_conn_sendmsg(msg, conn, NULL);
}
store classmark in vlr_subscr, not conn Store all Classmark information in the VLR. So, we now always know the Classmark 1 (mandatory IE for LU). This is visible in the msc_vlr_tests -- they no longer indicate "assuming A5/1 is supported" because classmark 1 is missing, because we now know the Classmark 1. Rationale: During Location Updating, we receive Classmark 1; during CM Service Request and Paging Response, we receive Classmark 2. So far we stored these only for the duration of the conn, so as soon as a LU is complete, we would forget CM1. In other words, for anything else than a LU Request, we had no Classmark 1 available at all. During Ciphering Mode Command, we rely on Classmark 1 to determine whether A5/1 is supported. That is moot if we don't even have a Classmark 1 for any CM Service Request or Paging Response initiated connections. The only reason that A5/1 worked is that we assume A5/1 to work if Classmark 1 is missing. To add to the confusion, if a phone indicated that it did *not* support A5/1 in the Classmark 1, according to spec we're supposed to not service it at all. A code comment however says that we instead want to heed the flag -- which so far was only present in a Location Updating initiated connection. Now we can make this decision without assuming things. This got my attention while hacking on sending a BSSMAP Classmark Request from the MSC if it finds missing Classmark information, and was surprised to see it it lacking CM1 to decide about A5/1. Change-Id: I27081bf6e9e017923b2d02607f7ea06beddad82a
2018-09-13 01:05:52 +00:00
static bool classmark1_is_r99(const struct gsm48_classmark1 *cm1)
{
return cm1->rev_lev >= 2;
}
static bool classmark2_is_r99(const uint8_t *cm2, uint8_t cm2_len)
{
uint8_t rev_lev;
if (!cm2_len)
return false;
rev_lev = (cm2[0] >> 5) & 0x3;
return rev_lev >= 2;
}
static bool classmark_is_r99(struct gsm_classmark *cm)
{
if (cm->classmark1_set)
store classmark in vlr_subscr, not conn Store all Classmark information in the VLR. So, we now always know the Classmark 1 (mandatory IE for LU). This is visible in the msc_vlr_tests -- they no longer indicate "assuming A5/1 is supported" because classmark 1 is missing, because we now know the Classmark 1. Rationale: During Location Updating, we receive Classmark 1; during CM Service Request and Paging Response, we receive Classmark 2. So far we stored these only for the duration of the conn, so as soon as a LU is complete, we would forget CM1. In other words, for anything else than a LU Request, we had no Classmark 1 available at all. During Ciphering Mode Command, we rely on Classmark 1 to determine whether A5/1 is supported. That is moot if we don't even have a Classmark 1 for any CM Service Request or Paging Response initiated connections. The only reason that A5/1 worked is that we assume A5/1 to work if Classmark 1 is missing. To add to the confusion, if a phone indicated that it did *not* support A5/1 in the Classmark 1, according to spec we're supposed to not service it at all. A code comment however says that we instead want to heed the flag -- which so far was only present in a Location Updating initiated connection. Now we can make this decision without assuming things. This got my attention while hacking on sending a BSSMAP Classmark Request from the MSC if it finds missing Classmark information, and was surprised to see it it lacking CM1 to decide about A5/1. Change-Id: I27081bf6e9e017923b2d02607f7ea06beddad82a
2018-09-13 01:05:52 +00:00
return classmark1_is_r99(&cm->classmark1);
return classmark2_is_r99(cm->classmark2, cm->classmark2_len);
}
static const char *classmark_a5_name(const struct gsm_classmark *cm)
{
static char buf[128];
char cm1[42];
char cm2[42];
char cm3[42];
if (cm->classmark1_set)
snprintf(cm1, sizeof(cm1), "cm1{a5/1=%s}",
cm->classmark1.a5_1 ? "not-supported":"supported" /* inverted logic */);
else
snprintf(cm1, sizeof(cm1), "no-cm1");
if (cm->classmark2_len >= 3)
snprintf(cm2, sizeof(cm2), " cm2{0x%x=%s%s}",
cm->classmark2[2],
cm->classmark2[2] & 0x1 ? " A5/2" : "",
cm->classmark2[2] & 0x2 ? " A5/3" : "");
else
snprintf(cm2, sizeof(cm2), " no-cm2");
if (cm->classmark3_len >= 1)
snprintf(cm3, sizeof(cm3), " cm3{0x%x=%s%s%s%s}",
cm->classmark3[0],
cm->classmark3[0] & (1 << 0) ? " A5/4" : "",
cm->classmark3[0] & (1 << 1) ? " A5/5" : "",
cm->classmark3[0] & (1 << 2) ? " A5/6" : "",
cm->classmark3[0] & (1 << 3) ? " A5/7" : "");
else
snprintf(cm3, sizeof(cm3), " no-cm3");
snprintf(buf, sizeof(buf), "%s%s%s", cm1, cm2, cm3);
return buf;
}
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
/* Determine if the given CLASSMARK (1/2/3) value permits a given A5/n cipher.
* Return 1 when the given A5/n is permitted, 0 when not, and negative if the respective MS CLASSMARK is
* not known, where the negative number indicates the classmark type: -2 means Classmark 2 is not
* available. */
static int classmark_supports_a5(const struct gsm_classmark *cm, uint8_t a5)
{
switch (a5) {
case 0:
/* all phones must implement A5/0, see 3GPP TS 43.020 4.9 */
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
return 1;
case 1:
/* 3GPP TS 43.020 4.9 requires A5/1 to be suppored by all phones and actually states:
* "The network shall not provide service to an MS which indicates that it does not
* support the ciphering algorithm A5/1.". However, let's be more tolerant based
* on policy here */
/* See 3GPP TS 24.008 10.5.1.7 */
if (!cm->classmark1_set) {
DEBUGP(DMSC, "CLASSMARK 1 unknown, assuming MS supports A5/1\n");
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
return -1;
} else {
if (cm->classmark1.a5_1)
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
return 0; /* Inverted logic for this bit! */
else
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
return 1;
}
break;
case 2:
case 3:
/* See 3GPP TS 24.008 10.5.1.6 */
if (cm->classmark2_len < 3) {
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
return -2;
} else {
if (cm->classmark2[2] & (1 << (a5-2)))
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
return 1;
else
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
return 0;
}
break;
case 4:
case 5:
case 6:
case 7:
/* See 3GPP TS 24.008 10.5.1.7 */
if (cm->classmark3_len < 1) {
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
return -3;
} else {
if (cm->classmark3[0] & (1 << (a5-4)))
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
return 1;
else
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
return 0;
}
break;
default:
return false;
}
}
int gsm48_conn_sendmsg(struct msgb *msg, struct ran_conn *conn, struct gsm_trans *trans)
{
struct gsm48_hdr *gh = (struct gsm48_hdr *) msg->data;
/* if we get passed a transaction reference, do some common
* work that the caller no longer has to do */
if (trans) {
gh->proto_discr = trans->protocol | (trans->transaction_id << 4);
OMSC_LINKID_CB(msg) = trans->dlci;
}
return msc_tx_dtap(conn, msg);
}
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
/* clear all transactions globally; used in case of MNCC socket disconnect */
void gsm0408_clear_all_trans(struct gsm_network *net, int protocol)
{
struct gsm_trans *trans, *temp;
LOGP(DCC, LOGL_NOTICE, "Clearing all currently active transactions!!!\n");
llist_for_each_entry_safe(trans, temp, &net->trans_list, entry) {
if (trans->protocol == protocol) {
trans->callref = 0;
trans_free(trans);
}
}
}
/* Chapter 9.2.14 : Send LOCATION UPDATING REJECT */
static int gsm0408_loc_upd_rej(struct ran_conn *conn, uint8_t cause)
{
struct msgb *msg;
msg = gsm48_create_loc_upd_rej(cause);
if (!msg) {
LOGP(DMM, LOGL_ERROR, "Failed to create msg for LOCATION UPDATING REJECT.\n");
return -1;
}
mscsplit: various preparations to separate MSC from BSC Disable large parts of the code that depend on BSC presence. The code sections disabled by #if BEFORE_MSCSPLIT shall be modified or dropped in the course of adding the A-interface. Don't set msg->lchan nor msg->dst. Don't use lchan in libmsc. Decouple lac from bts. Prepare entry/exit point for MSC -> BSC and MSC -> RNC communication: Add msc_ifaces.[hc], a_iface.c, with a general msc_tx_dtap() to redirect to different interfaces depending on the actual subscriber connection. While iu_tx() is going to be functional fairly soon, the a_tx() is going to be just a dummy for some time (see comment). Add Iu specific fields in gsm_subscriber_connection: the UE connection pointer and an indicator for the Integrity Protection status on Iu (to be fully implemented in later commits). Add lac member to gsm_subscriber_connection, to allow decoupling from bts->location_area_code. The conn->lac will actually be set in iu.c in an upcoming commit ("add iucs.[hc]"). move to libcommon-cs: gsm48_extract_mi(), gsm48_paging_extract_mi(). libmsc: duplicate gsm0808 / gsm48 functions (towards BSC). In osmo-nitb, libmsc would directly call the functions on the BSC level, not always via the bsc_api. When separating libmsc from libbsc, some functions are missing from the linkage. Hence duplicate these functions to libmsc, add an msc_ prefix for clarity, also add a _tx to gsm0808_cipher_mode(): * add msc_gsm0808_tx_cipher_mode() (dummy/stub) * add msc_gsm48_tx_mm_serv_ack() * add msc_gsm48_tx_mm_serv_rej() Call these from libmsc instead of * gsm0808_cipher_mode() * gsm48_tx_mm_serv_ack() * gsm48_tx_mm_serv_rej() Also add a comment related to msc_gsm0808_tx_cipher_mode() in two places. Remove internal RTP streaming code; OsmoNITB supported that, but for OsmoMSC, this will be done with an external MGCP gateway. Remove LCHAN_MODIFY from internal MNCC state machine. Temporarily disable all paging to be able to link libmsc without libbsc. Skip the paging part of channel_test because the paging is now disabled. Employ fake paging shims in order for msc_vlr_tests to still work. msc_compl_l3(): publish in .h, tweak return value. Use new libmsc enum values for return val, to avoid dependency on libbsc headers. Make callable from other scopes: publish in osmo_msc.h and remove 'static' in osmo_msc.c add gsm_encr to subscr_conn move subscr_request to gsm_subscriber.h subscr_request_channel() -> subscr_request_conn() move to libmsc: osmo_stats_vty_add_cmds() gsm_04_08: remove apply_codec_restrictions() gsm0408_test: use NULL for root ctx move to libbsc: gsm_bts_neighbor() move to libbsc: lchan_next_meas_rep() move vty config for t3212 to network level (periodic lu) remove unneccessary linking from some tests remove handle_abisip_signal() abis_rsl.c: don't use libvlr from libbsc gsm_subscriber_connection: put the LAC here, so that it is available without accessing conn->bts. In bsc_api.c, place this lac in conn for the sake of transition: Iu and A will use this new field to pass the LAC around, but in a completely separate OsmoBSC this is not actually needed. It can be removed again from osmo-bsc.git when the time has come. Siemens MRPCI: completely drop sending the MRPCI messages for now, they shall be added in osmo-bsc once the A-Interface code has settled. See OS#2389. Related: OS#1845 OS#2257 OS#2389 Change-Id: Id3705236350d5f69e447046b0a764bbabc3d493c
2017-05-08 13:12:20 +00:00
LOGP(DMM, LOGL_INFO, "Subscriber %s: LOCATION UPDATING REJECT\n",
vlr_subscr_name(conn->vsub));
return gsm48_conn_sendmsg(msg, conn, NULL);
}
/* Chapter 9.2.13 : Send LOCATION UPDATE ACCEPT */
static int gsm0408_loc_upd_acc(struct ran_conn *conn,
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
uint32_t send_tmsi)
{
struct msgb *msg = gsm48_msgb_alloc_name("GSM 04.08 LOC UPD ACC");
struct gsm48_hdr *gh;
struct gsm48_loc_area_id *lai;
uint8_t *mid;
struct osmo_location_area_id laid = {
.plmn = conn->network->plmn,
.lac = conn->lac,
};
gh = (struct gsm48_hdr *) msgb_put(msg, sizeof(*gh));
gh->proto_discr = GSM48_PDISC_MM;
gh->msg_type = GSM48_MT_MM_LOC_UPD_ACCEPT;
lai = (struct gsm48_loc_area_id *) msgb_put(msg, sizeof(*lai));
gsm48_generate_lai2(lai, &laid);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
if (send_tmsi == GSM_RESERVED_TMSI) {
/* we did not allocate a TMSI to the MS, so we need to
* include the IMSI in order for the MS to delete any
* old TMSI that might still be allocated */
uint8_t mi[10];
int len;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
len = gsm48_generate_mid_from_imsi(mi, conn->vsub->imsi);
mid = msgb_put(msg, len);
memcpy(mid, mi, len);
DEBUGP(DMM, "-> %s LOCATION UPDATE ACCEPT\n",
vlr_subscr_name(conn->vsub));
} else {
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
/* Include the TMSI, which means that the MS will send a
* TMSI REALLOCATION COMPLETE, and we should wait for
* that until T3250 expiration */
mid = msgb_put(msg, GSM48_MID_TMSI_LEN);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
gsm48_generate_mid_from_tmsi(mid, send_tmsi);
DEBUGP(DMM, "-> %s LOCATION UPDATE ACCEPT (TMSI = 0x%08x)\n",
vlr_subscr_name(conn->vsub),
send_tmsi);
}
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
/* TODO: Follow-on proceed */
/* TODO: CTS permission */
/* TODO: Equivalent PLMNs */
/* TODO: Emergency Number List */
/* TODO: Per-MS T3312 */
return gsm48_conn_sendmsg(msg, conn, NULL);
}
2009-02-03 12:59:45 +00:00
/* Transmit Chapter 9.2.10 Identity Request */
static int mm_tx_identity_req(struct ran_conn *conn, uint8_t id_type)
{
struct msgb *msg = gsm48_msgb_alloc_name("GSM 04.08 ID REQ");
struct gsm48_hdr *gh;
gh = (struct gsm48_hdr *) msgb_put(msg, sizeof(*gh) + 1);
gh->proto_discr = GSM48_PDISC_MM;
gh->msg_type = GSM48_MT_MM_ID_REQ;
gh->data[0] = id_type;
return gsm48_conn_sendmsg(msg, conn, NULL);
}
2009-02-03 12:59:45 +00:00
/* Parse Chapter 9.2.11 Identity Response */
static int mm_rx_id_resp(struct ran_conn *conn, struct msgb *msg)
{
struct gsm48_hdr *gh = msgb_l3(msg);
uint8_t *mi = gh->data+1;
uint8_t mi_len = gh->data[0];
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
if (!conn->vsub) {
LOGP(DMM, LOGL_ERROR,
"Rx MM Identity Response: invalid: no subscriber\n");
return -EINVAL;
}
DEBUGP(DMM, "IDENTITY RESPONSE: MI=%s\n", osmo_mi_name(mi, mi_len));
osmo_signal_dispatch(SS_SUBSCR, S_SUBSCR_IDENTITY, gh->data);
return vlr_subscr_rx_id_resp(conn->vsub, mi, mi_len);
}
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
/* Chapter 9.2.15: Receive Location Updating Request.
* Keep this function non-static for direct invocation by unit tests. */
int mm_rx_loc_upd_req(struct ran_conn *conn, struct msgb *msg)
{
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
struct gsm_network *net = conn->network;
struct gsm48_hdr *gh = msgb_l3(msg);
struct gsm48_loc_upd_req *lu;
uint8_t mi_type;
char mi_string[GSM48_MI_SIZE];
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
enum vlr_lu_type vlr_lu_type = VLR_LU_TYPE_REGULAR;
uint32_t tmsi;
char *imsi;
struct osmo_location_area_id old_lai, new_lai;
struct osmo_fsm_inst *lu_fsm;
bool is_utran;
lu = (struct gsm48_loc_upd_req *) gh->data;
if (ran_conn_is_establishing_auth_ciph(conn)) {
LOG_RAN_CONN_CAT(conn, DMM, LOGL_ERROR,
"Cannot accept another LU, conn already busy establishing authenticity;"
" extraneous LOCATION UPDATING REQUEST: MI=%s LU-type=%s\n",
osmo_mi_name(lu->mi, lu->mi_len), osmo_lu_type_name(lu->type));
refactor subscr_conn and subscr_conn_fsm de-/alloc Refactor: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. 2. Add separate AUTH_CIPH state to the FSM. 3. Use conn->use_count to trigger conn release. 4. Add separate RELEASING state to the FSM. 5. Add rate counters for each of the three Complete Layer 3 types. Details: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. Historically, a gsm_subscriber_connection was allocated in libbsc land, and only upon Complete Layer 3 did libmsc add the fsm instance. After splitting openbsc.git into a separate osmo-msc, this is no longer necessary, hence: Closely tie gsm_subscriber_connection allocation to the subscr_conn_fsm instance: talloc the conn as a child of the FSM instance, and discard the conn as soon as the FSM terminates. 2. Add separate AUTH_CIPH state to the FSM. Decoding the Complete Layer 3 message is distinctly separate from waiting for the VLR FSMs to conclude. Use the NEW state as "we don't know if this is a valid message yet", and the AUTH_CIPH state as "evaluating, don't release". A profound effect of this: should we for any odd reason fail to leave the FSM's NEW state, the conn will be released right at the end of msc_compl_l3(), without needing to trigger release in each code path. 3. Use conn->use_count to trigger conn release. Before, the FSM itself would hold a use count on the conn, and hence we would need to ask it whether it is ready to release the conn yet by dispatching events, to achieve a use_count decrement. Instead, unite the FSM instance and conn, and do not hold a use count by the FSM. Hence, trigger an FSM "UNUSED" event only when the use_count reaches zero. As long as use counts are done correctly, the FSM will terminate correctly. These exceptions: - The new AUTH_CIPH state explicitly ignores UNUSED events, since we expect the use count to reach zero while evaluating Authentication and Ciphering. (I experimented with holding a use count by AUTH_CIPH onenter() and releasing by onleave(), but the use count and thus the conn are released before the next state can initiate transactions that would increment the use count again. Same thing for the VLR FSMs holding a use count, they should be done before we advance to the next state. The easiest is to simply expect zero use count during the AUTH_CIPH state.) - A CM Service Request means that even though the MSC would be through with all it wants to do, we shall still wait for a request to follow from the MS. Hence the FSM holds a use count on itself while a CM Service is pending. - While waiting for a Release/Clear Complete, the FSM holds a use count on itself. 4. Add separate RELEASING state to the FSM. If we decide to release for other reasons than a use count reaching zero, we still need to be able to wait for the msc_dtap() use count on the conn to release. (An upcoming patch will further use the RELEASING state to properly wait for Clear Complete / Release Complete messages.) 5. Add rate counters for each of the three Complete Layer 3 types. Besides LU, also count CM Service Request and Paging Response acceptance/rejections. Without these counters, only very few of the auth+ciph outcomes actually show in the counters. Related: OS#3122 Change-Id: I55feb379e176a96a831e105b86202b17a0ffe889
2018-03-30 22:02:14 +00:00
return -EINVAL;
}
if (ran_conn_is_accepted(conn)) {
LOG_RAN_CONN_CAT(conn, DMM, LOGL_ERROR,
"Cannot accept another LU, conn already established;"
" extraneous LOCATION UPDATING REQUEST: MI=%s LU-type=%s\n",
osmo_mi_name(lu->mi, lu->mi_len), osmo_lu_type_name(lu->type));
refactor subscr_conn and subscr_conn_fsm de-/alloc Refactor: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. 2. Add separate AUTH_CIPH state to the FSM. 3. Use conn->use_count to trigger conn release. 4. Add separate RELEASING state to the FSM. 5. Add rate counters for each of the three Complete Layer 3 types. Details: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. Historically, a gsm_subscriber_connection was allocated in libbsc land, and only upon Complete Layer 3 did libmsc add the fsm instance. After splitting openbsc.git into a separate osmo-msc, this is no longer necessary, hence: Closely tie gsm_subscriber_connection allocation to the subscr_conn_fsm instance: talloc the conn as a child of the FSM instance, and discard the conn as soon as the FSM terminates. 2. Add separate AUTH_CIPH state to the FSM. Decoding the Complete Layer 3 message is distinctly separate from waiting for the VLR FSMs to conclude. Use the NEW state as "we don't know if this is a valid message yet", and the AUTH_CIPH state as "evaluating, don't release". A profound effect of this: should we for any odd reason fail to leave the FSM's NEW state, the conn will be released right at the end of msc_compl_l3(), without needing to trigger release in each code path. 3. Use conn->use_count to trigger conn release. Before, the FSM itself would hold a use count on the conn, and hence we would need to ask it whether it is ready to release the conn yet by dispatching events, to achieve a use_count decrement. Instead, unite the FSM instance and conn, and do not hold a use count by the FSM. Hence, trigger an FSM "UNUSED" event only when the use_count reaches zero. As long as use counts are done correctly, the FSM will terminate correctly. These exceptions: - The new AUTH_CIPH state explicitly ignores UNUSED events, since we expect the use count to reach zero while evaluating Authentication and Ciphering. (I experimented with holding a use count by AUTH_CIPH onenter() and releasing by onleave(), but the use count and thus the conn are released before the next state can initiate transactions that would increment the use count again. Same thing for the VLR FSMs holding a use count, they should be done before we advance to the next state. The easiest is to simply expect zero use count during the AUTH_CIPH state.) - A CM Service Request means that even though the MSC would be through with all it wants to do, we shall still wait for a request to follow from the MS. Hence the FSM holds a use count on itself while a CM Service is pending. - While waiting for a Release/Clear Complete, the FSM holds a use count on itself. 4. Add separate RELEASING state to the FSM. If we decide to release for other reasons than a use count reaching zero, we still need to be able to wait for the msc_dtap() use count on the conn to release. (An upcoming patch will further use the RELEASING state to properly wait for Clear Complete / Release Complete messages.) 5. Add rate counters for each of the three Complete Layer 3 types. Besides LU, also count CM Service Request and Paging Response acceptance/rejections. Without these counters, only very few of the auth+ciph outcomes actually show in the counters. Related: OS#3122 Change-Id: I55feb379e176a96a831e105b86202b17a0ffe889
2018-03-30 22:02:14 +00:00
return -EINVAL;
}
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
refactor log ctx for vlr_subscr and ran_conn ran_conn_get_conn_id(): instead of a talloc allocated string, return a static buffer in ran_conn_get_conn_id(). So far this function had no callers. Refactor ran_conn_update_id() API: during early L3-Complete, when no subscriber is associated yet, update the FSM Id by the MI type seen in the L3 Complete message: ran_conn_update_id_from_mi(). Later on set the vsub and re-update. Call vlr.ops->subscr_update when the TMSI is updated, so that log context includes the TMSI from then on. Enrich context for vlr_subscr_name and ran_conn fi name. Include all available information in vlr_subscr_name(); instead of either IMSI or MSISDN or TMSI, print all of them when present. Instead of a short log, rather have more valuable context. A context info would now look like: Process_Access_Request_VLR(IMSI-901700000014706:MSISDN-2023:TMSI-0x08BDE4EC:GERAN-A-3:PAGING_RESP) It does get quite long, but ensures easy correlation of any BSSAP / IuCS messages with log output, especially if multiple subscribers are busy at the same time. Print TMSI and TMSInew in uppercase hexadecimal, which is the typical representation in the telecom world. When showing the RAN conn id GERAN_A-00000017 becomes GERAN-A-23 - We usually write the conn_id in decimal. - Leading zeros are clutter and might suggest hexadecimal format. - 'GERAN-A' and 'UTRAN-Iu' are the strings defined by osmo_rat_type_name(). Depends: I7798c3ef983c2e333b2b9cbffef6f366f370bd81 (libosmocore) Depends: Ica25919758ef6cba8348da199b0ae7e0ba628798 (libosmocore) Change-Id: I66a68ce2eb8957a35855a3743d91a86299900834
2019-01-03 01:32:14 +00:00
conn->complete_layer3_type = COMPLETE_LAYER3_LU;
ran_conn_update_id_from_mi(conn, lu->mi, lu->mi_len);
LOG_RAN_CONN_CAT(conn, DMM, LOGL_DEBUG, "LOCATION UPDATING REQUEST: MI=%s LU-type=%s\n",
osmo_mi_name(lu->mi, lu->mi_len), osmo_lu_type_name(lu->type));
osmo_signal_dispatch(SS_SUBSCR, S_SUBSCR_IDENTITY, &lu->mi_len);
switch (lu->type) {
case GSM48_LUPD_NORMAL:
rate_ctr_inc(&conn->network->msc_ctrs->ctr[MSC_CTR_LOC_UPDATE_TYPE_NORMAL]);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vlr_lu_type = VLR_LU_TYPE_REGULAR;
break;
case GSM48_LUPD_IMSI_ATT:
rate_ctr_inc(&conn->network->msc_ctrs->ctr[MSC_CTR_LOC_UPDATE_TYPE_ATTACH]);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vlr_lu_type = VLR_LU_TYPE_IMSI_ATTACH;
break;
case GSM48_LUPD_PERIODIC:
rate_ctr_inc(&conn->network->msc_ctrs->ctr[MSC_CTR_LOC_UPDATE_TYPE_PERIODIC]);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vlr_lu_type = VLR_LU_TYPE_PERIODIC;
break;
}
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
/* TODO: 10.5.1.6 MS Classmark for UMTS / Classmark 2 */
/* TODO: 10.5.3.14 Aditional update parameters (CS fallback calls) */
/* TODO: 10.5.7.8 Device properties */
/* TODO: 10.5.1.15 MS network feature support */
refactor log ctx for vlr_subscr and ran_conn ran_conn_get_conn_id(): instead of a talloc allocated string, return a static buffer in ran_conn_get_conn_id(). So far this function had no callers. Refactor ran_conn_update_id() API: during early L3-Complete, when no subscriber is associated yet, update the FSM Id by the MI type seen in the L3 Complete message: ran_conn_update_id_from_mi(). Later on set the vsub and re-update. Call vlr.ops->subscr_update when the TMSI is updated, so that log context includes the TMSI from then on. Enrich context for vlr_subscr_name and ran_conn fi name. Include all available information in vlr_subscr_name(); instead of either IMSI or MSISDN or TMSI, print all of them when present. Instead of a short log, rather have more valuable context. A context info would now look like: Process_Access_Request_VLR(IMSI-901700000014706:MSISDN-2023:TMSI-0x08BDE4EC:GERAN-A-3:PAGING_RESP) It does get quite long, but ensures easy correlation of any BSSAP / IuCS messages with log output, especially if multiple subscribers are busy at the same time. Print TMSI and TMSInew in uppercase hexadecimal, which is the typical representation in the telecom world. When showing the RAN conn id GERAN_A-00000017 becomes GERAN-A-23 - We usually write the conn_id in decimal. - Leading zeros are clutter and might suggest hexadecimal format. - 'GERAN-A' and 'UTRAN-Iu' are the strings defined by osmo_rat_type_name(). Depends: I7798c3ef983c2e333b2b9cbffef6f366f370bd81 (libosmocore) Depends: Ica25919758ef6cba8348da199b0ae7e0ba628798 (libosmocore) Change-Id: I66a68ce2eb8957a35855a3743d91a86299900834
2019-01-03 01:32:14 +00:00
mi_type = lu->mi[0] & GSM_MI_TYPE_MASK;
gsm48_mi_to_string(mi_string, sizeof(mi_string), lu->mi, lu->mi_len);
switch (mi_type) {
case GSM_MI_TYPE_IMSI:
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
tmsi = GSM_RESERVED_TMSI;
imsi = mi_string;
2008-12-27 01:55:51 +00:00
break;
case GSM_MI_TYPE_TMSI:
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
tmsi = tmsi_from_string(mi_string);
imsi = NULL;
break;
default:
LOG_RAN_CONN_CAT(conn, DMM, LOGL_ERROR, "unknown mobile identity type\n");
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
tmsi = GSM_RESERVED_TMSI;
imsi = NULL;
break;
}
gsm48_decode_lai2(&lu->lai, &old_lai);
new_lai.plmn = conn->network->plmn;
mscsplit: various preparations to separate MSC from BSC Disable large parts of the code that depend on BSC presence. The code sections disabled by #if BEFORE_MSCSPLIT shall be modified or dropped in the course of adding the A-interface. Don't set msg->lchan nor msg->dst. Don't use lchan in libmsc. Decouple lac from bts. Prepare entry/exit point for MSC -> BSC and MSC -> RNC communication: Add msc_ifaces.[hc], a_iface.c, with a general msc_tx_dtap() to redirect to different interfaces depending on the actual subscriber connection. While iu_tx() is going to be functional fairly soon, the a_tx() is going to be just a dummy for some time (see comment). Add Iu specific fields in gsm_subscriber_connection: the UE connection pointer and an indicator for the Integrity Protection status on Iu (to be fully implemented in later commits). Add lac member to gsm_subscriber_connection, to allow decoupling from bts->location_area_code. The conn->lac will actually be set in iu.c in an upcoming commit ("add iucs.[hc]"). move to libcommon-cs: gsm48_extract_mi(), gsm48_paging_extract_mi(). libmsc: duplicate gsm0808 / gsm48 functions (towards BSC). In osmo-nitb, libmsc would directly call the functions on the BSC level, not always via the bsc_api. When separating libmsc from libbsc, some functions are missing from the linkage. Hence duplicate these functions to libmsc, add an msc_ prefix for clarity, also add a _tx to gsm0808_cipher_mode(): * add msc_gsm0808_tx_cipher_mode() (dummy/stub) * add msc_gsm48_tx_mm_serv_ack() * add msc_gsm48_tx_mm_serv_rej() Call these from libmsc instead of * gsm0808_cipher_mode() * gsm48_tx_mm_serv_ack() * gsm48_tx_mm_serv_rej() Also add a comment related to msc_gsm0808_tx_cipher_mode() in two places. Remove internal RTP streaming code; OsmoNITB supported that, but for OsmoMSC, this will be done with an external MGCP gateway. Remove LCHAN_MODIFY from internal MNCC state machine. Temporarily disable all paging to be able to link libmsc without libbsc. Skip the paging part of channel_test because the paging is now disabled. Employ fake paging shims in order for msc_vlr_tests to still work. msc_compl_l3(): publish in .h, tweak return value. Use new libmsc enum values for return val, to avoid dependency on libbsc headers. Make callable from other scopes: publish in osmo_msc.h and remove 'static' in osmo_msc.c add gsm_encr to subscr_conn move subscr_request to gsm_subscriber.h subscr_request_channel() -> subscr_request_conn() move to libmsc: osmo_stats_vty_add_cmds() gsm_04_08: remove apply_codec_restrictions() gsm0408_test: use NULL for root ctx move to libbsc: gsm_bts_neighbor() move to libbsc: lchan_next_meas_rep() move vty config for t3212 to network level (periodic lu) remove unneccessary linking from some tests remove handle_abisip_signal() abis_rsl.c: don't use libvlr from libbsc gsm_subscriber_connection: put the LAC here, so that it is available without accessing conn->bts. In bsc_api.c, place this lac in conn for the sake of transition: Iu and A will use this new field to pass the LAC around, but in a completely separate OsmoBSC this is not actually needed. It can be removed again from osmo-bsc.git when the time has come. Siemens MRPCI: completely drop sending the MRPCI messages for now, they shall be added in osmo-bsc once the A-Interface code has settled. See OS#2389. Related: OS#1845 OS#2257 OS#2389 Change-Id: Id3705236350d5f69e447046b0a764bbabc3d493c
2017-05-08 13:12:20 +00:00
new_lai.lac = conn->lac;
LOG_RAN_CONN_CAT(conn, DMM, LOGL_DEBUG, "LU/new-LAC: %u/%u\n", old_lai.lac, new_lai.lac);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
is_utran = (conn->via_ran == OSMO_RAT_UTRAN_IU);
lu_fsm = vlr_loc_update(conn->fi,
RAN_CONN_E_ACCEPTED, RAN_CONN_E_CN_CLOSE, NULL,
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
net->vlr, conn, vlr_lu_type, tmsi, imsi,
&old_lai, &new_lai,
is_utran || conn->network->authentication_required,
is_utran || conn->network->a5_encryption_mask > 0x01,
store classmark in vlr_subscr, not conn Store all Classmark information in the VLR. So, we now always know the Classmark 1 (mandatory IE for LU). This is visible in the msc_vlr_tests -- they no longer indicate "assuming A5/1 is supported" because classmark 1 is missing, because we now know the Classmark 1. Rationale: During Location Updating, we receive Classmark 1; during CM Service Request and Paging Response, we receive Classmark 2. So far we stored these only for the duration of the conn, so as soon as a LU is complete, we would forget CM1. In other words, for anything else than a LU Request, we had no Classmark 1 available at all. During Ciphering Mode Command, we rely on Classmark 1 to determine whether A5/1 is supported. That is moot if we don't even have a Classmark 1 for any CM Service Request or Paging Response initiated connections. The only reason that A5/1 worked is that we assume A5/1 to work if Classmark 1 is missing. To add to the confusion, if a phone indicated that it did *not* support A5/1 in the Classmark 1, according to spec we're supposed to not service it at all. A code comment however says that we instead want to heed the flag -- which so far was only present in a Location Updating initiated connection. Now we can make this decision without assuming things. This got my attention while hacking on sending a BSSMAP Classmark Request from the MSC if it finds missing Classmark information, and was surprised to see it it lacking CM1 to decide about A5/1. Change-Id: I27081bf6e9e017923b2d02607f7ea06beddad82a
2018-09-13 01:05:52 +00:00
classmark1_is_r99(&lu->classmark1),
is_utran,
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
net->vlr->cfg.assign_tmsi);
if (!lu_fsm) {
LOG_RAN_CONN(conn, LOGL_ERROR, "Can't start LU FSM\n");
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
return 0;
}
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
/* From vlr_loc_update() we expect an implicit dispatch of
* VLR_ULA_E_UPDATE_LA, and thus we expect msc_vlr_subscr_assoc() to
* already have been called and completed. Has an error occured? */
if (!conn->vsub || conn->vsub->lu_fsm != lu_fsm) {
LOG_RAN_CONN(conn, LOGL_ERROR, "internal error during Location Updating attempt\n");
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
return -EIO;
}
store classmark in vlr_subscr, not conn Store all Classmark information in the VLR. So, we now always know the Classmark 1 (mandatory IE for LU). This is visible in the msc_vlr_tests -- they no longer indicate "assuming A5/1 is supported" because classmark 1 is missing, because we now know the Classmark 1. Rationale: During Location Updating, we receive Classmark 1; during CM Service Request and Paging Response, we receive Classmark 2. So far we stored these only for the duration of the conn, so as soon as a LU is complete, we would forget CM1. In other words, for anything else than a LU Request, we had no Classmark 1 available at all. During Ciphering Mode Command, we rely on Classmark 1 to determine whether A5/1 is supported. That is moot if we don't even have a Classmark 1 for any CM Service Request or Paging Response initiated connections. The only reason that A5/1 worked is that we assume A5/1 to work if Classmark 1 is missing. To add to the confusion, if a phone indicated that it did *not* support A5/1 in the Classmark 1, according to spec we're supposed to not service it at all. A code comment however says that we instead want to heed the flag -- which so far was only present in a Location Updating initiated connection. Now we can make this decision without assuming things. This got my attention while hacking on sending a BSSMAP Classmark Request from the MSC if it finds missing Classmark information, and was surprised to see it it lacking CM1 to decide about A5/1. Change-Id: I27081bf6e9e017923b2d02607f7ea06beddad82a
2018-09-13 01:05:52 +00:00
conn->vsub->classmark.classmark1 = lu->classmark1;
conn->vsub->classmark.classmark1_set = true;
ran_conn_complete_layer_3(conn);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
return 0;
}
/* Turn int into semi-octet representation: 98 => 0x89 */
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
/* FIXME: libosmocore/libosmogsm */
static uint8_t bcdify(uint8_t value)
{
uint8_t ret;
ret = value / 10;
ret |= (value % 10) << 4;
return ret;
}
/* Section 9.2.15a */
int gsm48_tx_mm_info(struct ran_conn *conn)
{
struct msgb *msg = gsm48_msgb_alloc_name("GSM 04.08 MM INF");
struct gsm48_hdr *gh;
struct gsm_network *net = conn->network;
uint8_t *ptr8;
int name_len, name_pad;
time_t cur_t;
struct tm* gmt_time;
struct tm* local_time;
int tzunits;
int dst = 0;
gh = (struct gsm48_hdr *) msgb_put(msg, sizeof(*gh));
gh->proto_discr = GSM48_PDISC_MM;
gh->msg_type = GSM48_MT_MM_INFO;
if (net->name_long) {
#if 0
name_len = strlen(net->name_long);
/* 10.5.3.5a */
ptr8 = msgb_put(msg, 3);
ptr8[0] = GSM48_IE_NAME_LONG;
ptr8[1] = name_len*2 +1;
ptr8[2] = 0x90; /* UCS2, no spare bits, no CI */
ptr16 = (uint16_t *) msgb_put(msg, name_len*2);
for (i = 0; i < name_len; i++)
ptr16[i] = htons(net->name_long[i]);
/* FIXME: Use Cell Broadcast, not UCS-2, since
* UCS-2 is only supported by later revisions of the spec */
#endif
name_len = (strlen(net->name_long)*7)/8;
name_pad = (8 - strlen(net->name_long)*7)%8;
if (name_pad > 0)
name_len++;
/* 10.5.3.5a */
ptr8 = msgb_put(msg, 3);
ptr8[0] = GSM48_IE_NAME_LONG;
ptr8[1] = name_len +1;
ptr8[2] = 0x80 | name_pad; /* Cell Broadcast DCS, no CI */
ptr8 = msgb_put(msg, name_len);
gsm_7bit_encode_n(ptr8, name_len, net->name_long, NULL);
}
if (net->name_short) {
#if 0
name_len = strlen(net->name_short);
/* 10.5.3.5a */
ptr8 = (uint8_t *) msgb_put(msg, 3);
ptr8[0] = GSM48_IE_NAME_SHORT;
ptr8[1] = name_len*2 + 1;
ptr8[2] = 0x90; /* UCS2, no spare bits, no CI */
ptr16 = (uint16_t *) msgb_put(msg, name_len*2);
for (i = 0; i < name_len; i++)
ptr16[i] = htons(net->name_short[i]);
#endif
name_len = (strlen(net->name_short)*7)/8;
name_pad = (8 - strlen(net->name_short)*7)%8;
if (name_pad > 0)
name_len++;
/* 10.5.3.5a */
ptr8 = (uint8_t *) msgb_put(msg, 3);
ptr8[0] = GSM48_IE_NAME_SHORT;
ptr8[1] = name_len +1;
ptr8[2] = 0x80 | name_pad; /* Cell Broadcast DCS, no CI */
ptr8 = msgb_put(msg, name_len);
gsm_7bit_encode_n(ptr8, name_len, net->name_short, NULL);
}
/* Section 10.5.3.9 */
cur_t = time(NULL);
gmt_time = gmtime(&cur_t);
ptr8 = msgb_put(msg, 8);
ptr8[0] = GSM48_IE_NET_TIME_TZ;
ptr8[1] = bcdify(gmt_time->tm_year % 100);
ptr8[2] = bcdify(gmt_time->tm_mon + 1);
ptr8[3] = bcdify(gmt_time->tm_mday);
ptr8[4] = bcdify(gmt_time->tm_hour);
ptr8[5] = bcdify(gmt_time->tm_min);
ptr8[6] = bcdify(gmt_time->tm_sec);
if (net->tz.override) {
/* Convert tz.hr and tz.mn to units */
if (net->tz.hr < 0) {
tzunits = ((net->tz.hr/-1)*4);
tzunits = tzunits + (net->tz.mn/15);
ptr8[7] = bcdify(tzunits);
/* Set negative time */
ptr8[7] |= 0x08;
}
else {
tzunits = net->tz.hr*4;
tzunits = tzunits + (net->tz.mn/15);
ptr8[7] = bcdify(tzunits);
}
/* Convert DST value */
if (net->tz.dst >= 0 && net->tz.dst <= 2)
dst = net->tz.dst;
}
else {
/* Need to get GSM offset and convert into 15 min units */
/* This probably breaks if gmtoff returns a value not evenly divisible by 15? */
#ifdef HAVE_TM_GMTOFF_IN_TM
local_time = localtime(&cur_t);
tzunits = (local_time->tm_gmtoff/60)/15;
#else
/* find timezone offset */
time_t utc;
double offsetFromUTC;
utc = mktime(gmt_time);
local_time = localtime(&cur_t);
offsetFromUTC = difftime(cur_t, utc);
if (local_time->tm_isdst)
offsetFromUTC += 3600.0;
tzunits = ((int)offsetFromUTC) / 60 / 15;
#endif
if (tzunits < 0) {
tzunits = tzunits/-1;
ptr8[7] = bcdify(tzunits);
/* Flip it to negative */
ptr8[7] |= 0x08;
}
else
ptr8[7] = bcdify(tzunits);
/* Does not support DST +2 */
if (local_time->tm_isdst)
dst = 1;
}
if (dst) {
ptr8 = msgb_put(msg, 3);
ptr8[0] = GSM48_IE_NET_DST;
ptr8[1] = 1;
ptr8[2] = dst;
}
LOG_RAN_CONN(conn, LOGL_DEBUG, "Tx MM INFO\n");
return gsm48_conn_sendmsg(msg, conn, NULL);
}
/*! Send an Authentication Request to MS on the given RAN connection
* according to 3GPP/ETSI TS 24.008, Section 9.2.2.
* \param[in] conn Subscriber connection to send on.
* \param[in] rand Random challenge token to send, must be 16 bytes long.
* \param[in] autn r99: In case of UMTS mutual authentication, AUTN token to
* send; must be 16 bytes long, or pass NULL for plain GSM auth.
* \param[in] key_seq auth tuple's sequence number.
*/
int gsm48_tx_mm_auth_req(struct ran_conn *conn, uint8_t *rand,
uint8_t *autn, int key_seq)
{
struct msgb *msg = gsm48_msgb_alloc_name("GSM 04.08 AUTH REQ");
struct gsm48_hdr *gh = (struct gsm48_hdr *) msgb_put(msg, sizeof(*gh));
struct gsm48_auth_req *ar = (struct gsm48_auth_req *) msgb_put(msg, sizeof(*ar));
DEBUGP(DMM, "Tx AUTH REQ (rand = %s)\n", osmo_hexdump_nospc(rand, 16));
if (autn)
DEBUGP(DMM, " AUTH REQ (autn = %s)\n", osmo_hexdump_nospc(autn, 16));
gh->proto_discr = GSM48_PDISC_MM;
gh->msg_type = GSM48_MT_MM_AUTH_REQ;
ar->key_seq = key_seq;
/* 16 bytes RAND parameters */
osmo_static_assert(sizeof(ar->rand) == 16, sizeof_auth_req_r99_rand);
if (rand)
memcpy(ar->rand, rand, 16);
/* 16 bytes AUTN */
if (autn)
msgb_tlv_put(msg, GSM48_IE_AUTN, 16, autn);
return gsm48_conn_sendmsg(msg, conn, NULL);
}
/* Section 9.2.1 */
int gsm48_tx_mm_auth_rej(struct ran_conn *conn)
{
DEBUGP(DMM, "-> AUTH REJECT\n");
return gsm48_tx_simple(conn, GSM48_PDISC_MM, GSM48_MT_MM_AUTH_REJ);
}
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
static int msc_vlr_tx_cm_serv_acc(void *msc_conn_ref);
static int msc_vlr_tx_cm_serv_rej(void *msc_conn_ref, enum gsm48_reject_value result);
static int cm_serv_reuse_conn(struct ran_conn *conn, const uint8_t *mi_lv)
{
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
uint8_t mi_type;
char mi_string[GSM48_MI_SIZE];
uint32_t tmsi;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
gsm48_mi_to_string(mi_string, sizeof(mi_string), mi_lv+1, mi_lv[0]);
mi_type = mi_lv[1] & GSM_MI_TYPE_MASK;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
switch (mi_type) {
case GSM_MI_TYPE_IMSI:
if (vlr_subscr_matches_imsi(conn->vsub, mi_string))
goto accept_reuse;
break;
case GSM_MI_TYPE_TMSI:
tmsi = osmo_load32be(mi_lv+2);
if (vlr_subscr_matches_tmsi(conn->vsub, tmsi))
goto accept_reuse;
break;
case GSM_MI_TYPE_IMEI:
if (vlr_subscr_matches_imei(conn->vsub, mi_string))
goto accept_reuse;
break;
default:
break;
}
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
LOGP(DMM, LOGL_ERROR, "%s: CM Service Request with mismatching mobile identity: %s %s\n",
vlr_subscr_name(conn->vsub), gsm48_mi_type_name(mi_type), mi_string);
msc_vlr_tx_cm_serv_rej(conn, GSM48_REJECT_ILLEGAL_MS);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
return -EINVAL;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
accept_reuse:
DEBUGP(DMM, "%s: re-using already accepted connection\n",
vlr_subscr_name(conn->vsub));
refactor subscr_conn and subscr_conn_fsm de-/alloc Refactor: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. 2. Add separate AUTH_CIPH state to the FSM. 3. Use conn->use_count to trigger conn release. 4. Add separate RELEASING state to the FSM. 5. Add rate counters for each of the three Complete Layer 3 types. Details: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. Historically, a gsm_subscriber_connection was allocated in libbsc land, and only upon Complete Layer 3 did libmsc add the fsm instance. After splitting openbsc.git into a separate osmo-msc, this is no longer necessary, hence: Closely tie gsm_subscriber_connection allocation to the subscr_conn_fsm instance: talloc the conn as a child of the FSM instance, and discard the conn as soon as the FSM terminates. 2. Add separate AUTH_CIPH state to the FSM. Decoding the Complete Layer 3 message is distinctly separate from waiting for the VLR FSMs to conclude. Use the NEW state as "we don't know if this is a valid message yet", and the AUTH_CIPH state as "evaluating, don't release". A profound effect of this: should we for any odd reason fail to leave the FSM's NEW state, the conn will be released right at the end of msc_compl_l3(), without needing to trigger release in each code path. 3. Use conn->use_count to trigger conn release. Before, the FSM itself would hold a use count on the conn, and hence we would need to ask it whether it is ready to release the conn yet by dispatching events, to achieve a use_count decrement. Instead, unite the FSM instance and conn, and do not hold a use count by the FSM. Hence, trigger an FSM "UNUSED" event only when the use_count reaches zero. As long as use counts are done correctly, the FSM will terminate correctly. These exceptions: - The new AUTH_CIPH state explicitly ignores UNUSED events, since we expect the use count to reach zero while evaluating Authentication and Ciphering. (I experimented with holding a use count by AUTH_CIPH onenter() and releasing by onleave(), but the use count and thus the conn are released before the next state can initiate transactions that would increment the use count again. Same thing for the VLR FSMs holding a use count, they should be done before we advance to the next state. The easiest is to simply expect zero use count during the AUTH_CIPH state.) - A CM Service Request means that even though the MSC would be through with all it wants to do, we shall still wait for a request to follow from the MS. Hence the FSM holds a use count on itself while a CM Service is pending. - While waiting for a Release/Clear Complete, the FSM holds a use count on itself. 4. Add separate RELEASING state to the FSM. If we decide to release for other reasons than a use count reaching zero, we still need to be able to wait for the msc_dtap() use count on the conn to release. (An upcoming patch will further use the RELEASING state to properly wait for Clear Complete / Release Complete messages.) 5. Add rate counters for each of the three Complete Layer 3 types. Besides LU, also count CM Service Request and Paging Response acceptance/rejections. Without these counters, only very few of the auth+ciph outcomes actually show in the counters. Related: OS#3122 Change-Id: I55feb379e176a96a831e105b86202b17a0ffe889
2018-03-30 22:02:14 +00:00
if (!conn->received_cm_service_request) {
conn->received_cm_service_request = true;
ran_conn_get(conn, RAN_CONN_USE_CM_SERVICE);
refactor subscr_conn and subscr_conn_fsm de-/alloc Refactor: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. 2. Add separate AUTH_CIPH state to the FSM. 3. Use conn->use_count to trigger conn release. 4. Add separate RELEASING state to the FSM. 5. Add rate counters for each of the three Complete Layer 3 types. Details: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. Historically, a gsm_subscriber_connection was allocated in libbsc land, and only upon Complete Layer 3 did libmsc add the fsm instance. After splitting openbsc.git into a separate osmo-msc, this is no longer necessary, hence: Closely tie gsm_subscriber_connection allocation to the subscr_conn_fsm instance: talloc the conn as a child of the FSM instance, and discard the conn as soon as the FSM terminates. 2. Add separate AUTH_CIPH state to the FSM. Decoding the Complete Layer 3 message is distinctly separate from waiting for the VLR FSMs to conclude. Use the NEW state as "we don't know if this is a valid message yet", and the AUTH_CIPH state as "evaluating, don't release". A profound effect of this: should we for any odd reason fail to leave the FSM's NEW state, the conn will be released right at the end of msc_compl_l3(), without needing to trigger release in each code path. 3. Use conn->use_count to trigger conn release. Before, the FSM itself would hold a use count on the conn, and hence we would need to ask it whether it is ready to release the conn yet by dispatching events, to achieve a use_count decrement. Instead, unite the FSM instance and conn, and do not hold a use count by the FSM. Hence, trigger an FSM "UNUSED" event only when the use_count reaches zero. As long as use counts are done correctly, the FSM will terminate correctly. These exceptions: - The new AUTH_CIPH state explicitly ignores UNUSED events, since we expect the use count to reach zero while evaluating Authentication and Ciphering. (I experimented with holding a use count by AUTH_CIPH onenter() and releasing by onleave(), but the use count and thus the conn are released before the next state can initiate transactions that would increment the use count again. Same thing for the VLR FSMs holding a use count, they should be done before we advance to the next state. The easiest is to simply expect zero use count during the AUTH_CIPH state.) - A CM Service Request means that even though the MSC would be through with all it wants to do, we shall still wait for a request to follow from the MS. Hence the FSM holds a use count on itself while a CM Service is pending. - While waiting for a Release/Clear Complete, the FSM holds a use count on itself. 4. Add separate RELEASING state to the FSM. If we decide to release for other reasons than a use count reaching zero, we still need to be able to wait for the msc_dtap() use count on the conn to release. (An upcoming patch will further use the RELEASING state to properly wait for Clear Complete / Release Complete messages.) 5. Add rate counters for each of the three Complete Layer 3 types. Besides LU, also count CM Service Request and Paging Response acceptance/rejections. Without these counters, only very few of the auth+ciph outcomes actually show in the counters. Related: OS#3122 Change-Id: I55feb379e176a96a831e105b86202b17a0ffe889
2018-03-30 22:02:14 +00:00
}
refactor log ctx for vlr_subscr and ran_conn ran_conn_get_conn_id(): instead of a talloc allocated string, return a static buffer in ran_conn_get_conn_id(). So far this function had no callers. Refactor ran_conn_update_id() API: during early L3-Complete, when no subscriber is associated yet, update the FSM Id by the MI type seen in the L3 Complete message: ran_conn_update_id_from_mi(). Later on set the vsub and re-update. Call vlr.ops->subscr_update when the TMSI is updated, so that log context includes the TMSI from then on. Enrich context for vlr_subscr_name and ran_conn fi name. Include all available information in vlr_subscr_name(); instead of either IMSI or MSISDN or TMSI, print all of them when present. Instead of a short log, rather have more valuable context. A context info would now look like: Process_Access_Request_VLR(IMSI-901700000014706:MSISDN-2023:TMSI-0x08BDE4EC:GERAN-A-3:PAGING_RESP) It does get quite long, but ensures easy correlation of any BSSAP / IuCS messages with log output, especially if multiple subscribers are busy at the same time. Print TMSI and TMSInew in uppercase hexadecimal, which is the typical representation in the telecom world. When showing the RAN conn id GERAN_A-00000017 becomes GERAN-A-23 - We usually write the conn_id in decimal. - Leading zeros are clutter and might suggest hexadecimal format. - 'GERAN-A' and 'UTRAN-Iu' are the strings defined by osmo_rat_type_name(). Depends: I7798c3ef983c2e333b2b9cbffef6f366f370bd81 (libosmocore) Depends: Ica25919758ef6cba8348da199b0ae7e0ba628798 (libosmocore) Change-Id: I66a68ce2eb8957a35855a3743d91a86299900834
2019-01-03 01:32:14 +00:00
ran_conn_update_id(conn);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
return conn->network->vlr->ops.tx_cm_serv_acc(conn);
}
/*
* Handle CM Service Requests
* a) Verify that the packet is long enough to contain the information
* we require otherwsie reject with INCORRECT_MESSAGE
* b) Try to parse the TMSI. If we do not have one reject
* c) Check that we know the subscriber with the TMSI otherwise reject
* with a HLR cause
mscsplit: various preparations to separate MSC from BSC Disable large parts of the code that depend on BSC presence. The code sections disabled by #if BEFORE_MSCSPLIT shall be modified or dropped in the course of adding the A-interface. Don't set msg->lchan nor msg->dst. Don't use lchan in libmsc. Decouple lac from bts. Prepare entry/exit point for MSC -> BSC and MSC -> RNC communication: Add msc_ifaces.[hc], a_iface.c, with a general msc_tx_dtap() to redirect to different interfaces depending on the actual subscriber connection. While iu_tx() is going to be functional fairly soon, the a_tx() is going to be just a dummy for some time (see comment). Add Iu specific fields in gsm_subscriber_connection: the UE connection pointer and an indicator for the Integrity Protection status on Iu (to be fully implemented in later commits). Add lac member to gsm_subscriber_connection, to allow decoupling from bts->location_area_code. The conn->lac will actually be set in iu.c in an upcoming commit ("add iucs.[hc]"). move to libcommon-cs: gsm48_extract_mi(), gsm48_paging_extract_mi(). libmsc: duplicate gsm0808 / gsm48 functions (towards BSC). In osmo-nitb, libmsc would directly call the functions on the BSC level, not always via the bsc_api. When separating libmsc from libbsc, some functions are missing from the linkage. Hence duplicate these functions to libmsc, add an msc_ prefix for clarity, also add a _tx to gsm0808_cipher_mode(): * add msc_gsm0808_tx_cipher_mode() (dummy/stub) * add msc_gsm48_tx_mm_serv_ack() * add msc_gsm48_tx_mm_serv_rej() Call these from libmsc instead of * gsm0808_cipher_mode() * gsm48_tx_mm_serv_ack() * gsm48_tx_mm_serv_rej() Also add a comment related to msc_gsm0808_tx_cipher_mode() in two places. Remove internal RTP streaming code; OsmoNITB supported that, but for OsmoMSC, this will be done with an external MGCP gateway. Remove LCHAN_MODIFY from internal MNCC state machine. Temporarily disable all paging to be able to link libmsc without libbsc. Skip the paging part of channel_test because the paging is now disabled. Employ fake paging shims in order for msc_vlr_tests to still work. msc_compl_l3(): publish in .h, tweak return value. Use new libmsc enum values for return val, to avoid dependency on libbsc headers. Make callable from other scopes: publish in osmo_msc.h and remove 'static' in osmo_msc.c add gsm_encr to subscr_conn move subscr_request to gsm_subscriber.h subscr_request_channel() -> subscr_request_conn() move to libmsc: osmo_stats_vty_add_cmds() gsm_04_08: remove apply_codec_restrictions() gsm0408_test: use NULL for root ctx move to libbsc: gsm_bts_neighbor() move to libbsc: lchan_next_meas_rep() move vty config for t3212 to network level (periodic lu) remove unneccessary linking from some tests remove handle_abisip_signal() abis_rsl.c: don't use libvlr from libbsc gsm_subscriber_connection: put the LAC here, so that it is available without accessing conn->bts. In bsc_api.c, place this lac in conn for the sake of transition: Iu and A will use this new field to pass the LAC around, but in a completely separate OsmoBSC this is not actually needed. It can be removed again from osmo-bsc.git when the time has come. Siemens MRPCI: completely drop sending the MRPCI messages for now, they shall be added in osmo-bsc once the A-Interface code has settled. See OS#2389. Related: OS#1845 OS#2257 OS#2389 Change-Id: Id3705236350d5f69e447046b0a764bbabc3d493c
2017-05-08 13:12:20 +00:00
* d) Set the subscriber on the conn and accept
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
*
* Keep this function non-static for direct invocation by unit tests.
*/
int gsm48_rx_mm_serv_req(struct ran_conn *conn, struct msgb *msg)
2008-12-27 01:55:51 +00:00
{
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
struct gsm_network *net = conn->network;
uint8_t mi_type;
2008-12-27 01:55:51 +00:00
struct gsm48_hdr *gh = msgb_l3(msg);
struct gsm48_service_request *req =
(struct gsm48_service_request *)gh->data;
2010-05-14 00:02:08 +00:00
/* unfortunately in Phase1 the classmark2 length is variable */
uint8_t classmark2_len = gh->data[1];
uint8_t *classmark2 = gh->data+2;
refactor subscr_conn and subscr_conn_fsm de-/alloc Refactor: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. 2. Add separate AUTH_CIPH state to the FSM. 3. Use conn->use_count to trigger conn release. 4. Add separate RELEASING state to the FSM. 5. Add rate counters for each of the three Complete Layer 3 types. Details: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. Historically, a gsm_subscriber_connection was allocated in libbsc land, and only upon Complete Layer 3 did libmsc add the fsm instance. After splitting openbsc.git into a separate osmo-msc, this is no longer necessary, hence: Closely tie gsm_subscriber_connection allocation to the subscr_conn_fsm instance: talloc the conn as a child of the FSM instance, and discard the conn as soon as the FSM terminates. 2. Add separate AUTH_CIPH state to the FSM. Decoding the Complete Layer 3 message is distinctly separate from waiting for the VLR FSMs to conclude. Use the NEW state as "we don't know if this is a valid message yet", and the AUTH_CIPH state as "evaluating, don't release". A profound effect of this: should we for any odd reason fail to leave the FSM's NEW state, the conn will be released right at the end of msc_compl_l3(), without needing to trigger release in each code path. 3. Use conn->use_count to trigger conn release. Before, the FSM itself would hold a use count on the conn, and hence we would need to ask it whether it is ready to release the conn yet by dispatching events, to achieve a use_count decrement. Instead, unite the FSM instance and conn, and do not hold a use count by the FSM. Hence, trigger an FSM "UNUSED" event only when the use_count reaches zero. As long as use counts are done correctly, the FSM will terminate correctly. These exceptions: - The new AUTH_CIPH state explicitly ignores UNUSED events, since we expect the use count to reach zero while evaluating Authentication and Ciphering. (I experimented with holding a use count by AUTH_CIPH onenter() and releasing by onleave(), but the use count and thus the conn are released before the next state can initiate transactions that would increment the use count again. Same thing for the VLR FSMs holding a use count, they should be done before we advance to the next state. The easiest is to simply expect zero use count during the AUTH_CIPH state.) - A CM Service Request means that even though the MSC would be through with all it wants to do, we shall still wait for a request to follow from the MS. Hence the FSM holds a use count on itself while a CM Service is pending. - While waiting for a Release/Clear Complete, the FSM holds a use count on itself. 4. Add separate RELEASING state to the FSM. If we decide to release for other reasons than a use count reaching zero, we still need to be able to wait for the msc_dtap() use count on the conn to release. (An upcoming patch will further use the RELEASING state to properly wait for Clear Complete / Release Complete messages.) 5. Add rate counters for each of the three Complete Layer 3 types. Besides LU, also count CM Service Request and Paging Response acceptance/rejections. Without these counters, only very few of the auth+ciph outcomes actually show in the counters. Related: OS#3122 Change-Id: I55feb379e176a96a831e105b86202b17a0ffe889
2018-03-30 22:02:14 +00:00
uint8_t *mi_p = classmark2 + classmark2_len;
uint8_t mi_len = *mi_p;
uint8_t *mi = mi_p + 1;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
struct osmo_location_area_id lai;
bool is_utran;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
lai.plmn = conn->network->plmn;
mscsplit: various preparations to separate MSC from BSC Disable large parts of the code that depend on BSC presence. The code sections disabled by #if BEFORE_MSCSPLIT shall be modified or dropped in the course of adding the A-interface. Don't set msg->lchan nor msg->dst. Don't use lchan in libmsc. Decouple lac from bts. Prepare entry/exit point for MSC -> BSC and MSC -> RNC communication: Add msc_ifaces.[hc], a_iface.c, with a general msc_tx_dtap() to redirect to different interfaces depending on the actual subscriber connection. While iu_tx() is going to be functional fairly soon, the a_tx() is going to be just a dummy for some time (see comment). Add Iu specific fields in gsm_subscriber_connection: the UE connection pointer and an indicator for the Integrity Protection status on Iu (to be fully implemented in later commits). Add lac member to gsm_subscriber_connection, to allow decoupling from bts->location_area_code. The conn->lac will actually be set in iu.c in an upcoming commit ("add iucs.[hc]"). move to libcommon-cs: gsm48_extract_mi(), gsm48_paging_extract_mi(). libmsc: duplicate gsm0808 / gsm48 functions (towards BSC). In osmo-nitb, libmsc would directly call the functions on the BSC level, not always via the bsc_api. When separating libmsc from libbsc, some functions are missing from the linkage. Hence duplicate these functions to libmsc, add an msc_ prefix for clarity, also add a _tx to gsm0808_cipher_mode(): * add msc_gsm0808_tx_cipher_mode() (dummy/stub) * add msc_gsm48_tx_mm_serv_ack() * add msc_gsm48_tx_mm_serv_rej() Call these from libmsc instead of * gsm0808_cipher_mode() * gsm48_tx_mm_serv_ack() * gsm48_tx_mm_serv_rej() Also add a comment related to msc_gsm0808_tx_cipher_mode() in two places. Remove internal RTP streaming code; OsmoNITB supported that, but for OsmoMSC, this will be done with an external MGCP gateway. Remove LCHAN_MODIFY from internal MNCC state machine. Temporarily disable all paging to be able to link libmsc without libbsc. Skip the paging part of channel_test because the paging is now disabled. Employ fake paging shims in order for msc_vlr_tests to still work. msc_compl_l3(): publish in .h, tweak return value. Use new libmsc enum values for return val, to avoid dependency on libbsc headers. Make callable from other scopes: publish in osmo_msc.h and remove 'static' in osmo_msc.c add gsm_encr to subscr_conn move subscr_request to gsm_subscriber.h subscr_request_channel() -> subscr_request_conn() move to libmsc: osmo_stats_vty_add_cmds() gsm_04_08: remove apply_codec_restrictions() gsm0408_test: use NULL for root ctx move to libbsc: gsm_bts_neighbor() move to libbsc: lchan_next_meas_rep() move vty config for t3212 to network level (periodic lu) remove unneccessary linking from some tests remove handle_abisip_signal() abis_rsl.c: don't use libvlr from libbsc gsm_subscriber_connection: put the LAC here, so that it is available without accessing conn->bts. In bsc_api.c, place this lac in conn for the sake of transition: Iu and A will use this new field to pass the LAC around, but in a completely separate OsmoBSC this is not actually needed. It can be removed again from osmo-bsc.git when the time has come. Siemens MRPCI: completely drop sending the MRPCI messages for now, they shall be added in osmo-bsc once the A-Interface code has settled. See OS#2389. Related: OS#1845 OS#2257 OS#2389 Change-Id: Id3705236350d5f69e447046b0a764bbabc3d493c
2017-05-08 13:12:20 +00:00
lai.lac = conn->lac;
if (msg->data_len < sizeof(struct gsm48_service_request*)) {
LOG_RAN_CONN(conn, LOGL_ERROR, "Rx CM SERVICE REQUEST: wrong message size (%u < %zu)\n",
msg->data_len, sizeof(struct gsm48_service_request*));
mscsplit: various preparations to separate MSC from BSC Disable large parts of the code that depend on BSC presence. The code sections disabled by #if BEFORE_MSCSPLIT shall be modified or dropped in the course of adding the A-interface. Don't set msg->lchan nor msg->dst. Don't use lchan in libmsc. Decouple lac from bts. Prepare entry/exit point for MSC -> BSC and MSC -> RNC communication: Add msc_ifaces.[hc], a_iface.c, with a general msc_tx_dtap() to redirect to different interfaces depending on the actual subscriber connection. While iu_tx() is going to be functional fairly soon, the a_tx() is going to be just a dummy for some time (see comment). Add Iu specific fields in gsm_subscriber_connection: the UE connection pointer and an indicator for the Integrity Protection status on Iu (to be fully implemented in later commits). Add lac member to gsm_subscriber_connection, to allow decoupling from bts->location_area_code. The conn->lac will actually be set in iu.c in an upcoming commit ("add iucs.[hc]"). move to libcommon-cs: gsm48_extract_mi(), gsm48_paging_extract_mi(). libmsc: duplicate gsm0808 / gsm48 functions (towards BSC). In osmo-nitb, libmsc would directly call the functions on the BSC level, not always via the bsc_api. When separating libmsc from libbsc, some functions are missing from the linkage. Hence duplicate these functions to libmsc, add an msc_ prefix for clarity, also add a _tx to gsm0808_cipher_mode(): * add msc_gsm0808_tx_cipher_mode() (dummy/stub) * add msc_gsm48_tx_mm_serv_ack() * add msc_gsm48_tx_mm_serv_rej() Call these from libmsc instead of * gsm0808_cipher_mode() * gsm48_tx_mm_serv_ack() * gsm48_tx_mm_serv_rej() Also add a comment related to msc_gsm0808_tx_cipher_mode() in two places. Remove internal RTP streaming code; OsmoNITB supported that, but for OsmoMSC, this will be done with an external MGCP gateway. Remove LCHAN_MODIFY from internal MNCC state machine. Temporarily disable all paging to be able to link libmsc without libbsc. Skip the paging part of channel_test because the paging is now disabled. Employ fake paging shims in order for msc_vlr_tests to still work. msc_compl_l3(): publish in .h, tweak return value. Use new libmsc enum values for return val, to avoid dependency on libbsc headers. Make callable from other scopes: publish in osmo_msc.h and remove 'static' in osmo_msc.c add gsm_encr to subscr_conn move subscr_request to gsm_subscriber.h subscr_request_channel() -> subscr_request_conn() move to libmsc: osmo_stats_vty_add_cmds() gsm_04_08: remove apply_codec_restrictions() gsm0408_test: use NULL for root ctx move to libbsc: gsm_bts_neighbor() move to libbsc: lchan_next_meas_rep() move vty config for t3212 to network level (periodic lu) remove unneccessary linking from some tests remove handle_abisip_signal() abis_rsl.c: don't use libvlr from libbsc gsm_subscriber_connection: put the LAC here, so that it is available without accessing conn->bts. In bsc_api.c, place this lac in conn for the sake of transition: Iu and A will use this new field to pass the LAC around, but in a completely separate OsmoBSC this is not actually needed. It can be removed again from osmo-bsc.git when the time has come. Siemens MRPCI: completely drop sending the MRPCI messages for now, they shall be added in osmo-bsc once the A-Interface code has settled. See OS#2389. Related: OS#1845 OS#2257 OS#2389 Change-Id: Id3705236350d5f69e447046b0a764bbabc3d493c
2017-05-08 13:12:20 +00:00
return msc_gsm48_tx_mm_serv_rej(conn,
GSM48_REJECT_INCORRECT_MESSAGE);
}
if (msg->data_len < req->mi_len + 6) {
LOG_RAN_CONN(conn, LOGL_ERROR, "Rx CM SERVICE REQUEST: message does not fit in packet\n");
mscsplit: various preparations to separate MSC from BSC Disable large parts of the code that depend on BSC presence. The code sections disabled by #if BEFORE_MSCSPLIT shall be modified or dropped in the course of adding the A-interface. Don't set msg->lchan nor msg->dst. Don't use lchan in libmsc. Decouple lac from bts. Prepare entry/exit point for MSC -> BSC and MSC -> RNC communication: Add msc_ifaces.[hc], a_iface.c, with a general msc_tx_dtap() to redirect to different interfaces depending on the actual subscriber connection. While iu_tx() is going to be functional fairly soon, the a_tx() is going to be just a dummy for some time (see comment). Add Iu specific fields in gsm_subscriber_connection: the UE connection pointer and an indicator for the Integrity Protection status on Iu (to be fully implemented in later commits). Add lac member to gsm_subscriber_connection, to allow decoupling from bts->location_area_code. The conn->lac will actually be set in iu.c in an upcoming commit ("add iucs.[hc]"). move to libcommon-cs: gsm48_extract_mi(), gsm48_paging_extract_mi(). libmsc: duplicate gsm0808 / gsm48 functions (towards BSC). In osmo-nitb, libmsc would directly call the functions on the BSC level, not always via the bsc_api. When separating libmsc from libbsc, some functions are missing from the linkage. Hence duplicate these functions to libmsc, add an msc_ prefix for clarity, also add a _tx to gsm0808_cipher_mode(): * add msc_gsm0808_tx_cipher_mode() (dummy/stub) * add msc_gsm48_tx_mm_serv_ack() * add msc_gsm48_tx_mm_serv_rej() Call these from libmsc instead of * gsm0808_cipher_mode() * gsm48_tx_mm_serv_ack() * gsm48_tx_mm_serv_rej() Also add a comment related to msc_gsm0808_tx_cipher_mode() in two places. Remove internal RTP streaming code; OsmoNITB supported that, but for OsmoMSC, this will be done with an external MGCP gateway. Remove LCHAN_MODIFY from internal MNCC state machine. Temporarily disable all paging to be able to link libmsc without libbsc. Skip the paging part of channel_test because the paging is now disabled. Employ fake paging shims in order for msc_vlr_tests to still work. msc_compl_l3(): publish in .h, tweak return value. Use new libmsc enum values for return val, to avoid dependency on libbsc headers. Make callable from other scopes: publish in osmo_msc.h and remove 'static' in osmo_msc.c add gsm_encr to subscr_conn move subscr_request to gsm_subscriber.h subscr_request_channel() -> subscr_request_conn() move to libmsc: osmo_stats_vty_add_cmds() gsm_04_08: remove apply_codec_restrictions() gsm0408_test: use NULL for root ctx move to libbsc: gsm_bts_neighbor() move to libbsc: lchan_next_meas_rep() move vty config for t3212 to network level (periodic lu) remove unneccessary linking from some tests remove handle_abisip_signal() abis_rsl.c: don't use libvlr from libbsc gsm_subscriber_connection: put the LAC here, so that it is available without accessing conn->bts. In bsc_api.c, place this lac in conn for the sake of transition: Iu and A will use this new field to pass the LAC around, but in a completely separate OsmoBSC this is not actually needed. It can be removed again from osmo-bsc.git when the time has come. Siemens MRPCI: completely drop sending the MRPCI messages for now, they shall be added in osmo-bsc once the A-Interface code has settled. See OS#2389. Related: OS#1845 OS#2257 OS#2389 Change-Id: Id3705236350d5f69e447046b0a764bbabc3d493c
2017-05-08 13:12:20 +00:00
return msc_gsm48_tx_mm_serv_rej(conn,
GSM48_REJECT_INCORRECT_MESSAGE);
}
if (ran_conn_is_establishing_auth_ciph(conn)) {
LOG_RAN_CONN(conn, LOGL_ERROR,
"Cannot accept CM Service Request, conn already busy establishing authenticity\n");
msc_vlr_tx_cm_serv_rej(conn, GSM48_REJECT_CONGESTION);
return -EINVAL;
/* or should we accept and note down the service request anyway? */
}
refactor log ctx for vlr_subscr and ran_conn ran_conn_get_conn_id(): instead of a talloc allocated string, return a static buffer in ran_conn_get_conn_id(). So far this function had no callers. Refactor ran_conn_update_id() API: during early L3-Complete, when no subscriber is associated yet, update the FSM Id by the MI type seen in the L3 Complete message: ran_conn_update_id_from_mi(). Later on set the vsub and re-update. Call vlr.ops->subscr_update when the TMSI is updated, so that log context includes the TMSI from then on. Enrich context for vlr_subscr_name and ran_conn fi name. Include all available information in vlr_subscr_name(); instead of either IMSI or MSISDN or TMSI, print all of them when present. Instead of a short log, rather have more valuable context. A context info would now look like: Process_Access_Request_VLR(IMSI-901700000014706:MSISDN-2023:TMSI-0x08BDE4EC:GERAN-A-3:PAGING_RESP) It does get quite long, but ensures easy correlation of any BSSAP / IuCS messages with log output, especially if multiple subscribers are busy at the same time. Print TMSI and TMSInew in uppercase hexadecimal, which is the typical representation in the telecom world. When showing the RAN conn id GERAN_A-00000017 becomes GERAN-A-23 - We usually write the conn_id in decimal. - Leading zeros are clutter and might suggest hexadecimal format. - 'GERAN-A' and 'UTRAN-Iu' are the strings defined by osmo_rat_type_name(). Depends: I7798c3ef983c2e333b2b9cbffef6f366f370bd81 (libosmocore) Depends: Ica25919758ef6cba8348da199b0ae7e0ba628798 (libosmocore) Change-Id: I66a68ce2eb8957a35855a3743d91a86299900834
2019-01-03 01:32:14 +00:00
conn->complete_layer3_type = COMPLETE_LAYER3_CM_SERVICE_REQ;
ran_conn_update_id_from_mi(conn, mi, mi_len);
LOG_RAN_CONN_CAT(conn, DMM, LOGL_DEBUG, "Rx CM SERVICE REQUEST cm_service_type=0x%02x\n",
req->cm_service_type);
refactor log ctx for vlr_subscr and ran_conn ran_conn_get_conn_id(): instead of a talloc allocated string, return a static buffer in ran_conn_get_conn_id(). So far this function had no callers. Refactor ran_conn_update_id() API: during early L3-Complete, when no subscriber is associated yet, update the FSM Id by the MI type seen in the L3 Complete message: ran_conn_update_id_from_mi(). Later on set the vsub and re-update. Call vlr.ops->subscr_update when the TMSI is updated, so that log context includes the TMSI from then on. Enrich context for vlr_subscr_name and ran_conn fi name. Include all available information in vlr_subscr_name(); instead of either IMSI or MSISDN or TMSI, print all of them when present. Instead of a short log, rather have more valuable context. A context info would now look like: Process_Access_Request_VLR(IMSI-901700000014706:MSISDN-2023:TMSI-0x08BDE4EC:GERAN-A-3:PAGING_RESP) It does get quite long, but ensures easy correlation of any BSSAP / IuCS messages with log output, especially if multiple subscribers are busy at the same time. Print TMSI and TMSInew in uppercase hexadecimal, which is the typical representation in the telecom world. When showing the RAN conn id GERAN_A-00000017 becomes GERAN-A-23 - We usually write the conn_id in decimal. - Leading zeros are clutter and might suggest hexadecimal format. - 'GERAN-A' and 'UTRAN-Iu' are the strings defined by osmo_rat_type_name(). Depends: I7798c3ef983c2e333b2b9cbffef6f366f370bd81 (libosmocore) Depends: Ica25919758ef6cba8348da199b0ae7e0ba628798 (libosmocore) Change-Id: I66a68ce2eb8957a35855a3743d91a86299900834
2019-01-03 01:32:14 +00:00
mi_type = (mi && mi_len) ? (mi[0] & GSM_MI_TYPE_MASK) : GSM_MI_TYPE_NONE;
switch (mi_type) {
case GSM_MI_TYPE_IMSI:
case GSM_MI_TYPE_TMSI:
/* continue below */
break;
case GSM_MI_TYPE_IMEI:
if (req->cm_service_type == GSM48_CMSERV_EMERGENCY) {
/* We don't do emergency calls by IMEI */
LOG_RAN_CONN(conn, LOGL_NOTICE, "Tx CM SERVICE REQUEST REJECT\n");
return msc_gsm48_tx_mm_serv_rej(conn, GSM48_REJECT_IMEI_NOT_ACCEPTED);
}
/* fall-through for non-emergency setup */
default:
refactor log ctx for vlr_subscr and ran_conn ran_conn_get_conn_id(): instead of a talloc allocated string, return a static buffer in ran_conn_get_conn_id(). So far this function had no callers. Refactor ran_conn_update_id() API: during early L3-Complete, when no subscriber is associated yet, update the FSM Id by the MI type seen in the L3 Complete message: ran_conn_update_id_from_mi(). Later on set the vsub and re-update. Call vlr.ops->subscr_update when the TMSI is updated, so that log context includes the TMSI from then on. Enrich context for vlr_subscr_name and ran_conn fi name. Include all available information in vlr_subscr_name(); instead of either IMSI or MSISDN or TMSI, print all of them when present. Instead of a short log, rather have more valuable context. A context info would now look like: Process_Access_Request_VLR(IMSI-901700000014706:MSISDN-2023:TMSI-0x08BDE4EC:GERAN-A-3:PAGING_RESP) It does get quite long, but ensures easy correlation of any BSSAP / IuCS messages with log output, especially if multiple subscribers are busy at the same time. Print TMSI and TMSInew in uppercase hexadecimal, which is the typical representation in the telecom world. When showing the RAN conn id GERAN_A-00000017 becomes GERAN-A-23 - We usually write the conn_id in decimal. - Leading zeros are clutter and might suggest hexadecimal format. - 'GERAN-A' and 'UTRAN-Iu' are the strings defined by osmo_rat_type_name(). Depends: I7798c3ef983c2e333b2b9cbffef6f366f370bd81 (libosmocore) Depends: Ica25919758ef6cba8348da199b0ae7e0ba628798 (libosmocore) Change-Id: I66a68ce2eb8957a35855a3743d91a86299900834
2019-01-03 01:32:14 +00:00
LOG_RAN_CONN(conn, LOGL_ERROR, "MI type is not expected: %s\n", gsm48_mi_type_name(mi_type));
mscsplit: various preparations to separate MSC from BSC Disable large parts of the code that depend on BSC presence. The code sections disabled by #if BEFORE_MSCSPLIT shall be modified or dropped in the course of adding the A-interface. Don't set msg->lchan nor msg->dst. Don't use lchan in libmsc. Decouple lac from bts. Prepare entry/exit point for MSC -> BSC and MSC -> RNC communication: Add msc_ifaces.[hc], a_iface.c, with a general msc_tx_dtap() to redirect to different interfaces depending on the actual subscriber connection. While iu_tx() is going to be functional fairly soon, the a_tx() is going to be just a dummy for some time (see comment). Add Iu specific fields in gsm_subscriber_connection: the UE connection pointer and an indicator for the Integrity Protection status on Iu (to be fully implemented in later commits). Add lac member to gsm_subscriber_connection, to allow decoupling from bts->location_area_code. The conn->lac will actually be set in iu.c in an upcoming commit ("add iucs.[hc]"). move to libcommon-cs: gsm48_extract_mi(), gsm48_paging_extract_mi(). libmsc: duplicate gsm0808 / gsm48 functions (towards BSC). In osmo-nitb, libmsc would directly call the functions on the BSC level, not always via the bsc_api. When separating libmsc from libbsc, some functions are missing from the linkage. Hence duplicate these functions to libmsc, add an msc_ prefix for clarity, also add a _tx to gsm0808_cipher_mode(): * add msc_gsm0808_tx_cipher_mode() (dummy/stub) * add msc_gsm48_tx_mm_serv_ack() * add msc_gsm48_tx_mm_serv_rej() Call these from libmsc instead of * gsm0808_cipher_mode() * gsm48_tx_mm_serv_ack() * gsm48_tx_mm_serv_rej() Also add a comment related to msc_gsm0808_tx_cipher_mode() in two places. Remove internal RTP streaming code; OsmoNITB supported that, but for OsmoMSC, this will be done with an external MGCP gateway. Remove LCHAN_MODIFY from internal MNCC state machine. Temporarily disable all paging to be able to link libmsc without libbsc. Skip the paging part of channel_test because the paging is now disabled. Employ fake paging shims in order for msc_vlr_tests to still work. msc_compl_l3(): publish in .h, tweak return value. Use new libmsc enum values for return val, to avoid dependency on libbsc headers. Make callable from other scopes: publish in osmo_msc.h and remove 'static' in osmo_msc.c add gsm_encr to subscr_conn move subscr_request to gsm_subscriber.h subscr_request_channel() -> subscr_request_conn() move to libmsc: osmo_stats_vty_add_cmds() gsm_04_08: remove apply_codec_restrictions() gsm0408_test: use NULL for root ctx move to libbsc: gsm_bts_neighbor() move to libbsc: lchan_next_meas_rep() move vty config for t3212 to network level (periodic lu) remove unneccessary linking from some tests remove handle_abisip_signal() abis_rsl.c: don't use libvlr from libbsc gsm_subscriber_connection: put the LAC here, so that it is available without accessing conn->bts. In bsc_api.c, place this lac in conn for the sake of transition: Iu and A will use this new field to pass the LAC around, but in a completely separate OsmoBSC this is not actually needed. It can be removed again from osmo-bsc.git when the time has come. Siemens MRPCI: completely drop sending the MRPCI messages for now, they shall be added in osmo-bsc once the A-Interface code has settled. See OS#2389. Related: OS#1845 OS#2257 OS#2389 Change-Id: Id3705236350d5f69e447046b0a764bbabc3d493c
2017-05-08 13:12:20 +00:00
return msc_gsm48_tx_mm_serv_rej(conn,
GSM48_REJECT_INCORRECT_MESSAGE);
}
2008-12-27 21:45:37 +00:00
switch (req->cm_service_type) {
case GSM48_CMSERV_MO_CALL_PACKET:
case GSM48_CMSERV_EMERGENCY:
case GSM48_CMSERV_SMS:
case GSM48_CMSERV_SUP_SERV:
/* continue below */
break;
default:
return msc_gsm48_tx_mm_serv_rej(conn, GSM48_REJECT_SRV_OPT_NOT_SUPPORTED);
}
if (ran_conn_is_accepted(conn))
refactor subscr_conn and subscr_conn_fsm de-/alloc Refactor: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. 2. Add separate AUTH_CIPH state to the FSM. 3. Use conn->use_count to trigger conn release. 4. Add separate RELEASING state to the FSM. 5. Add rate counters for each of the three Complete Layer 3 types. Details: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. Historically, a gsm_subscriber_connection was allocated in libbsc land, and only upon Complete Layer 3 did libmsc add the fsm instance. After splitting openbsc.git into a separate osmo-msc, this is no longer necessary, hence: Closely tie gsm_subscriber_connection allocation to the subscr_conn_fsm instance: talloc the conn as a child of the FSM instance, and discard the conn as soon as the FSM terminates. 2. Add separate AUTH_CIPH state to the FSM. Decoding the Complete Layer 3 message is distinctly separate from waiting for the VLR FSMs to conclude. Use the NEW state as "we don't know if this is a valid message yet", and the AUTH_CIPH state as "evaluating, don't release". A profound effect of this: should we for any odd reason fail to leave the FSM's NEW state, the conn will be released right at the end of msc_compl_l3(), without needing to trigger release in each code path. 3. Use conn->use_count to trigger conn release. Before, the FSM itself would hold a use count on the conn, and hence we would need to ask it whether it is ready to release the conn yet by dispatching events, to achieve a use_count decrement. Instead, unite the FSM instance and conn, and do not hold a use count by the FSM. Hence, trigger an FSM "UNUSED" event only when the use_count reaches zero. As long as use counts are done correctly, the FSM will terminate correctly. These exceptions: - The new AUTH_CIPH state explicitly ignores UNUSED events, since we expect the use count to reach zero while evaluating Authentication and Ciphering. (I experimented with holding a use count by AUTH_CIPH onenter() and releasing by onleave(), but the use count and thus the conn are released before the next state can initiate transactions that would increment the use count again. Same thing for the VLR FSMs holding a use count, they should be done before we advance to the next state. The easiest is to simply expect zero use count during the AUTH_CIPH state.) - A CM Service Request means that even though the MSC would be through with all it wants to do, we shall still wait for a request to follow from the MS. Hence the FSM holds a use count on itself while a CM Service is pending. - While waiting for a Release/Clear Complete, the FSM holds a use count on itself. 4. Add separate RELEASING state to the FSM. If we decide to release for other reasons than a use count reaching zero, we still need to be able to wait for the msc_dtap() use count on the conn to release. (An upcoming patch will further use the RELEASING state to properly wait for Clear Complete / Release Complete messages.) 5. Add rate counters for each of the three Complete Layer 3 types. Besides LU, also count CM Service Request and Paging Response acceptance/rejections. Without these counters, only very few of the auth+ciph outcomes actually show in the counters. Related: OS#3122 Change-Id: I55feb379e176a96a831e105b86202b17a0ffe889
2018-03-30 22:02:14 +00:00
return cm_serv_reuse_conn(conn, mi_p);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
refactor subscr_conn and subscr_conn_fsm de-/alloc Refactor: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. 2. Add separate AUTH_CIPH state to the FSM. 3. Use conn->use_count to trigger conn release. 4. Add separate RELEASING state to the FSM. 5. Add rate counters for each of the three Complete Layer 3 types. Details: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. Historically, a gsm_subscriber_connection was allocated in libbsc land, and only upon Complete Layer 3 did libmsc add the fsm instance. After splitting openbsc.git into a separate osmo-msc, this is no longer necessary, hence: Closely tie gsm_subscriber_connection allocation to the subscr_conn_fsm instance: talloc the conn as a child of the FSM instance, and discard the conn as soon as the FSM terminates. 2. Add separate AUTH_CIPH state to the FSM. Decoding the Complete Layer 3 message is distinctly separate from waiting for the VLR FSMs to conclude. Use the NEW state as "we don't know if this is a valid message yet", and the AUTH_CIPH state as "evaluating, don't release". A profound effect of this: should we for any odd reason fail to leave the FSM's NEW state, the conn will be released right at the end of msc_compl_l3(), without needing to trigger release in each code path. 3. Use conn->use_count to trigger conn release. Before, the FSM itself would hold a use count on the conn, and hence we would need to ask it whether it is ready to release the conn yet by dispatching events, to achieve a use_count decrement. Instead, unite the FSM instance and conn, and do not hold a use count by the FSM. Hence, trigger an FSM "UNUSED" event only when the use_count reaches zero. As long as use counts are done correctly, the FSM will terminate correctly. These exceptions: - The new AUTH_CIPH state explicitly ignores UNUSED events, since we expect the use count to reach zero while evaluating Authentication and Ciphering. (I experimented with holding a use count by AUTH_CIPH onenter() and releasing by onleave(), but the use count and thus the conn are released before the next state can initiate transactions that would increment the use count again. Same thing for the VLR FSMs holding a use count, they should be done before we advance to the next state. The easiest is to simply expect zero use count during the AUTH_CIPH state.) - A CM Service Request means that even though the MSC would be through with all it wants to do, we shall still wait for a request to follow from the MS. Hence the FSM holds a use count on itself while a CM Service is pending. - While waiting for a Release/Clear Complete, the FSM holds a use count on itself. 4. Add separate RELEASING state to the FSM. If we decide to release for other reasons than a use count reaching zero, we still need to be able to wait for the msc_dtap() use count on the conn to release. (An upcoming patch will further use the RELEASING state to properly wait for Clear Complete / Release Complete messages.) 5. Add rate counters for each of the three Complete Layer 3 types. Besides LU, also count CM Service Request and Paging Response acceptance/rejections. Without these counters, only very few of the auth+ciph outcomes actually show in the counters. Related: OS#3122 Change-Id: I55feb379e176a96a831e105b86202b17a0ffe889
2018-03-30 22:02:14 +00:00
osmo_signal_dispatch(SS_SUBSCR, S_SUBSCR_IDENTITY, mi_p);
is_utran = (conn->via_ran == OSMO_RAT_UTRAN_IU);
vlr_proc_acc_req(conn->fi,
RAN_CONN_E_ACCEPTED, RAN_CONN_E_CN_CLOSE, NULL,
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
net->vlr, conn,
VLR_PR_ARQ_T_CM_SERV_REQ, mi-1, &lai,
is_utran || conn->network->authentication_required,
is_utran || conn->network->a5_encryption_mask > 0x01,
store classmark in vlr_subscr, not conn Store all Classmark information in the VLR. So, we now always know the Classmark 1 (mandatory IE for LU). This is visible in the msc_vlr_tests -- they no longer indicate "assuming A5/1 is supported" because classmark 1 is missing, because we now know the Classmark 1. Rationale: During Location Updating, we receive Classmark 1; during CM Service Request and Paging Response, we receive Classmark 2. So far we stored these only for the duration of the conn, so as soon as a LU is complete, we would forget CM1. In other words, for anything else than a LU Request, we had no Classmark 1 available at all. During Ciphering Mode Command, we rely on Classmark 1 to determine whether A5/1 is supported. That is moot if we don't even have a Classmark 1 for any CM Service Request or Paging Response initiated connections. The only reason that A5/1 worked is that we assume A5/1 to work if Classmark 1 is missing. To add to the confusion, if a phone indicated that it did *not* support A5/1 in the Classmark 1, according to spec we're supposed to not service it at all. A code comment however says that we instead want to heed the flag -- which so far was only present in a Location Updating initiated connection. Now we can make this decision without assuming things. This got my attention while hacking on sending a BSSMAP Classmark Request from the MSC if it finds missing Classmark information, and was surprised to see it it lacking CM1 to decide about A5/1. Change-Id: I27081bf6e9e017923b2d02607f7ea06beddad82a
2018-09-13 01:05:52 +00:00
classmark2_is_r99(classmark2, classmark2_len),
is_utran);
store classmark in vlr_subscr, not conn Store all Classmark information in the VLR. So, we now always know the Classmark 1 (mandatory IE for LU). This is visible in the msc_vlr_tests -- they no longer indicate "assuming A5/1 is supported" because classmark 1 is missing, because we now know the Classmark 1. Rationale: During Location Updating, we receive Classmark 1; during CM Service Request and Paging Response, we receive Classmark 2. So far we stored these only for the duration of the conn, so as soon as a LU is complete, we would forget CM1. In other words, for anything else than a LU Request, we had no Classmark 1 available at all. During Ciphering Mode Command, we rely on Classmark 1 to determine whether A5/1 is supported. That is moot if we don't even have a Classmark 1 for any CM Service Request or Paging Response initiated connections. The only reason that A5/1 worked is that we assume A5/1 to work if Classmark 1 is missing. To add to the confusion, if a phone indicated that it did *not* support A5/1 in the Classmark 1, according to spec we're supposed to not service it at all. A code comment however says that we instead want to heed the flag -- which so far was only present in a Location Updating initiated connection. Now we can make this decision without assuming things. This got my attention while hacking on sending a BSSMAP Classmark Request from the MSC if it finds missing Classmark information, and was surprised to see it it lacking CM1 to decide about A5/1. Change-Id: I27081bf6e9e017923b2d02607f7ea06beddad82a
2018-09-13 01:05:52 +00:00
/* From vlr_proc_acc_req() we expect an implicit dispatch of PR_ARQ_E_START we expect
* msc_vlr_subscr_assoc() to already have been called and completed. Has an error occured? */
if (!conn->vsub) {
LOG_RAN_CONN(conn, LOGL_ERROR, "subscriber not allowed to do a CM Service Request\n");
store classmark in vlr_subscr, not conn Store all Classmark information in the VLR. So, we now always know the Classmark 1 (mandatory IE for LU). This is visible in the msc_vlr_tests -- they no longer indicate "assuming A5/1 is supported" because classmark 1 is missing, because we now know the Classmark 1. Rationale: During Location Updating, we receive Classmark 1; during CM Service Request and Paging Response, we receive Classmark 2. So far we stored these only for the duration of the conn, so as soon as a LU is complete, we would forget CM1. In other words, for anything else than a LU Request, we had no Classmark 1 available at all. During Ciphering Mode Command, we rely on Classmark 1 to determine whether A5/1 is supported. That is moot if we don't even have a Classmark 1 for any CM Service Request or Paging Response initiated connections. The only reason that A5/1 worked is that we assume A5/1 to work if Classmark 1 is missing. To add to the confusion, if a phone indicated that it did *not* support A5/1 in the Classmark 1, according to spec we're supposed to not service it at all. A code comment however says that we instead want to heed the flag -- which so far was only present in a Location Updating initiated connection. Now we can make this decision without assuming things. This got my attention while hacking on sending a BSSMAP Classmark Request from the MSC if it finds missing Classmark information, and was surprised to see it it lacking CM1 to decide about A5/1. Change-Id: I27081bf6e9e017923b2d02607f7ea06beddad82a
2018-09-13 01:05:52 +00:00
return -EIO;
}
memcpy(conn->vsub->classmark.classmark2, classmark2, classmark2_len);
conn->vsub->classmark.classmark2_len = classmark2_len;
ran_conn_complete_layer_3(conn);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
return 0;
2008-12-27 01:55:51 +00:00
}
/* Receive a CM Re-establish Request */
static int gsm48_rx_cm_reest_req(struct ran_conn *conn, struct msgb *msg)
{
uint8_t mi_type;
char mi_string[GSM48_MI_SIZE];
struct gsm48_hdr *gh = msgb_l3(msg);
uint8_t classmark2_len = gh->data[1];
uint8_t *classmark2 = gh->data+2;
uint8_t mi_len = *(classmark2 + classmark2_len);
uint8_t *mi = (classmark2 + classmark2_len + 1);
gsm48_mi_to_string(mi_string, sizeof(mi_string), mi, mi_len);
mi_type = mi[0] & GSM_MI_TYPE_MASK;
DEBUGP(DMM, "<- CM RE-ESTABLISH REQUEST MI(%s)=%s\n", gsm48_mi_type_name(mi_type), mi_string);
/* we don't support CM call re-establishment */
return msc_gsm48_tx_mm_serv_rej(conn, GSM48_REJECT_SRV_OPT_NOT_SUPPORTED);
}
static int gsm48_rx_mm_imsi_detach_ind(struct ran_conn *conn, struct msgb *msg)
2009-02-22 21:14:55 +00:00
{
struct gsm_network *network = conn->network;
2009-02-22 21:14:55 +00:00
struct gsm48_hdr *gh = msgb_l3(msg);
struct gsm48_imsi_detach_ind *idi =
(struct gsm48_imsi_detach_ind *) gh->data;
uint8_t mi_type = idi->mi[0] & GSM_MI_TYPE_MASK;
char mi_string[GSM48_MI_SIZE];
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
struct vlr_subscr *vsub = NULL;
2009-02-22 21:14:55 +00:00
gsm48_mi_to_string(mi_string, sizeof(mi_string), idi->mi, idi->mi_len);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
DEBUGP(DMM, "IMSI DETACH INDICATION: MI(%s)=%s\n",
gsm48_mi_type_name(mi_type), mi_string);
2009-02-22 21:14:55 +00:00
rate_ctr_inc(&network->msc_ctrs->ctr[MSC_CTR_LOC_UPDATE_TYPE_DETACH]);
2009-02-22 21:14:55 +00:00
switch (mi_type) {
case GSM_MI_TYPE_TMSI:
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vsub = vlr_subscr_find_by_tmsi(network->vlr,
tmsi_from_string(mi_string));
2009-02-22 21:14:55 +00:00
break;
case GSM_MI_TYPE_IMSI:
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vsub = vlr_subscr_find_by_imsi(network->vlr, mi_string);
2009-02-22 21:14:55 +00:00
break;
case GSM_MI_TYPE_IMEI:
case GSM_MI_TYPE_IMEISV:
/* no sim card... FIXME: what to do ? */
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
LOGP(DMM, LOGL_ERROR, "MI(%s)=%s: unimplemented mobile identity type\n",
gsm48_mi_type_name(mi_type), mi_string);
2009-02-22 21:14:55 +00:00
break;
default:
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
LOGP(DMM, LOGL_ERROR, "MI(%s)=%s: unknown mobile identity type\n",
gsm48_mi_type_name(mi_type), mi_string);
2009-02-22 21:14:55 +00:00
break;
}
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
if (!vsub) {
LOGP(DMM, LOGL_ERROR, "IMSI DETACH for unknown subscriber MI(%s)=%s\n",
gsm48_mi_type_name(mi_type), mi_string);
} else {
LOGP(DMM, LOGL_INFO, "IMSI DETACH for %s\n", vlr_subscr_name(vsub));
if (vsub->cs.is_paging)
subscr_paging_cancel(vsub, GSM_PAGING_EXPIRED);
store classmark in vlr_subscr, not conn Store all Classmark information in the VLR. So, we now always know the Classmark 1 (mandatory IE for LU). This is visible in the msc_vlr_tests -- they no longer indicate "assuming A5/1 is supported" because classmark 1 is missing, because we now know the Classmark 1. Rationale: During Location Updating, we receive Classmark 1; during CM Service Request and Paging Response, we receive Classmark 2. So far we stored these only for the duration of the conn, so as soon as a LU is complete, we would forget CM1. In other words, for anything else than a LU Request, we had no Classmark 1 available at all. During Ciphering Mode Command, we rely on Classmark 1 to determine whether A5/1 is supported. That is moot if we don't even have a Classmark 1 for any CM Service Request or Paging Response initiated connections. The only reason that A5/1 worked is that we assume A5/1 to work if Classmark 1 is missing. To add to the confusion, if a phone indicated that it did *not* support A5/1 in the Classmark 1, according to spec we're supposed to not service it at all. A code comment however says that we instead want to heed the flag -- which so far was only present in a Location Updating initiated connection. Now we can make this decision without assuming things. This got my attention while hacking on sending a BSSMAP Classmark Request from the MSC if it finds missing Classmark information, and was surprised to see it it lacking CM1 to decide about A5/1. Change-Id: I27081bf6e9e017923b2d02607f7ea06beddad82a
2018-09-13 01:05:52 +00:00
/* We already got Classmark 1 during Location Updating ... but well, ok */
vsub->classmark.classmark1 = idi->classmark1;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vlr_subscr_rx_imsi_detach(vsub);
osmo_signal_dispatch(SS_SUBSCR, S_SUBSCR_DETACHED, vsub);
vlr_subscr_put(vsub);
}
ran_conn_close(conn, 0);
2009-02-22 21:14:55 +00:00
return 0;
}
static int gsm48_rx_mm_status(struct msgb *msg)
{
struct gsm48_hdr *gh = msgb_l3(msg);
DEBUGP(DMM, "MM STATUS (reject cause 0x%02x)\n", gh->data[0]);
return 0;
}
static int parse_gsm_auth_resp(uint8_t *res, uint8_t *res_len,
struct ran_conn *conn,
struct msgb *msg)
{
struct gsm48_hdr *gh = msgb_l3(msg);
struct gsm48_auth_resp *ar = (struct gsm48_auth_resp*) gh->data;
if (msgb_l3len(msg) < sizeof(*gh) + sizeof(*ar)) {
LOGP(DMM, LOGL_ERROR,
"%s: MM AUTHENTICATION RESPONSE:"
" l3 length invalid: %u\n",
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vlr_subscr_name(conn->vsub), msgb_l3len(msg));
return -EINVAL;
}
*res_len = sizeof(ar->sres);
memcpy(res, ar->sres, sizeof(ar->sres));
return 0;
}
static int parse_umts_auth_resp(uint8_t *res, uint8_t *res_len,
struct ran_conn *conn,
struct msgb *msg)
{
struct gsm48_hdr *gh;
uint8_t *data;
uint8_t iei;
uint8_t ie_len;
unsigned int data_len;
/* First parse the GSM part */
if (parse_gsm_auth_resp(res, res_len, conn, msg))
return -EINVAL;
OSMO_ASSERT(*res_len == 4);
/* Then add the extended res part */
gh = msgb_l3(msg);
data = gh->data + sizeof(struct gsm48_auth_resp);
data_len = msgb_l3len(msg) - (data - (uint8_t*)msgb_l3(msg));
if (data_len < 3) {
LOGP(DMM, LOGL_ERROR,
"%s: MM AUTHENTICATION RESPONSE:"
" l3 length invalid: %u\n",
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vlr_subscr_name(conn->vsub), msgb_l3len(msg));
return -EINVAL;
}
iei = data[0];
ie_len = data[1];
if (iei != GSM48_IE_AUTH_RES_EXT) {
LOGP(DMM, LOGL_ERROR,
"%s: MM R99 AUTHENTICATION RESPONSE:"
" expected IEI 0x%02x, got 0x%02x\n",
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vlr_subscr_name(conn->vsub),
GSM48_IE_AUTH_RES_EXT, iei);
return -EINVAL;
}
if (ie_len > 12) {
LOGP(DMM, LOGL_ERROR,
"%s: MM R99 AUTHENTICATION RESPONSE:"
" extended Auth Resp IE 0x%02x is too large: %u bytes\n",
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vlr_subscr_name(conn->vsub), GSM48_IE_AUTH_RES_EXT, ie_len);
return -EINVAL;
}
*res_len += ie_len;
memcpy(res + 4, &data[2], ie_len);
return 0;
}
/* Chapter 9.2.3: Authentication Response */
static int gsm48_rx_mm_auth_resp(struct ran_conn *conn, struct msgb *msg)
{
uint8_t res[16];
uint8_t res_len;
int rc;
bool is_umts;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
if (!conn->vsub) {
LOGP(DMM, LOGL_ERROR,
"MM AUTHENTICATION RESPONSE: invalid: no subscriber\n");
ran_conn_close(conn, GSM_CAUSE_AUTH_FAILED);
return -EINVAL;
}
is_umts = (msgb_l3len(msg) > sizeof(struct gsm48_hdr) + sizeof(struct gsm48_auth_resp));
if (is_umts)
rc = parse_umts_auth_resp(res, &res_len, conn, msg);
else
rc = parse_gsm_auth_resp(res, &res_len, conn, msg);
if (rc) {
LOGP(DMM, LOGL_ERROR,
"%s: MM AUTHENTICATION RESPONSE: invalid: parsing %s AKA Auth Response"
" failed with rc=%d; dispatching zero length SRES/RES to trigger failure\n",
vlr_subscr_name(conn->vsub), is_umts ? "UMTS" : "GSM", rc);
memset(res, 0, sizeof(res));
res_len = 0;
}
DEBUGP(DMM, "%s: MM %s AUTHENTICATION RESPONSE (%s = %s)\n",
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vlr_subscr_name(conn->vsub),
is_umts ? "UMTS" : "GSM", is_umts ? "res" : "sres",
osmo_hexdump_nospc(res, res_len));
store classmark in vlr_subscr, not conn Store all Classmark information in the VLR. So, we now always know the Classmark 1 (mandatory IE for LU). This is visible in the msc_vlr_tests -- they no longer indicate "assuming A5/1 is supported" because classmark 1 is missing, because we now know the Classmark 1. Rationale: During Location Updating, we receive Classmark 1; during CM Service Request and Paging Response, we receive Classmark 2. So far we stored these only for the duration of the conn, so as soon as a LU is complete, we would forget CM1. In other words, for anything else than a LU Request, we had no Classmark 1 available at all. During Ciphering Mode Command, we rely on Classmark 1 to determine whether A5/1 is supported. That is moot if we don't even have a Classmark 1 for any CM Service Request or Paging Response initiated connections. The only reason that A5/1 worked is that we assume A5/1 to work if Classmark 1 is missing. To add to the confusion, if a phone indicated that it did *not* support A5/1 in the Classmark 1, according to spec we're supposed to not service it at all. A code comment however says that we instead want to heed the flag -- which so far was only present in a Location Updating initiated connection. Now we can make this decision without assuming things. This got my attention while hacking on sending a BSSMAP Classmark Request from the MSC if it finds missing Classmark information, and was surprised to see it it lacking CM1 to decide about A5/1. Change-Id: I27081bf6e9e017923b2d02607f7ea06beddad82a
2018-09-13 01:05:52 +00:00
return vlr_subscr_rx_auth_resp(conn->vsub, classmark_is_r99(&conn->vsub->classmark),
conn->via_ran == OSMO_RAT_UTRAN_IU,
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
res, res_len);
}
static int gsm48_rx_mm_auth_fail(struct ran_conn *conn, struct msgb *msg)
{
struct gsm48_hdr *gh = msgb_l3(msg);
uint8_t cause;
uint8_t auts_tag;
uint8_t auts_len;
uint8_t *auts;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
if (!conn->vsub) {
LOGP(DMM, LOGL_ERROR,
"MM R99 AUTHENTICATION FAILURE: invalid: no subscriber\n");
ran_conn_close(conn, GSM_CAUSE_AUTH_FAILED);
return -EINVAL;
}
if (msgb_l3len(msg) < sizeof(*gh) + 1) {
LOGP(DMM, LOGL_ERROR,
"%s: MM R99 AUTHENTICATION FAILURE:"
" l3 length invalid: %u\n",
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vlr_subscr_name(conn->vsub), msgb_l3len(msg));
ran_conn_close(conn, GSM_CAUSE_AUTH_FAILED);
return -EINVAL;
}
cause = gh->data[0];
if (cause != GSM48_REJECT_SYNCH_FAILURE) {
LOGP(DMM, LOGL_INFO,
"%s: MM R99 AUTHENTICATION FAILURE: cause 0x%0x\n",
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vlr_subscr_name(conn->vsub), cause);
vlr_subscr_rx_auth_fail(conn->vsub, NULL);
return 0;
}
/* This is a Synch Failure procedure, which should pass an AUTS to
* resynchronize the sequence nr with the HLR. Expecting exactly one
* TLV with 14 bytes of AUTS. */
if (msgb_l3len(msg) < sizeof(*gh) + 1 + 2) {
LOGP(DMM, LOGL_INFO,
"%s: MM R99 AUTHENTICATION FAILURE:"
" invalid Synch Failure: missing AUTS IE\n",
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vlr_subscr_name(conn->vsub));
ran_conn_close(conn, GSM_CAUSE_AUTH_FAILED);
return -EINVAL;
}
auts_tag = gh->data[1];
auts_len = gh->data[2];
auts = &gh->data[3];
if (auts_tag != GSM48_IE_AUTS
|| auts_len != 14) {
LOGP(DMM, LOGL_INFO,
"%s: MM R99 AUTHENTICATION FAILURE:"
" invalid Synch Failure:"
" expected AUTS IE 0x%02x of 14 bytes,"
" got IE 0x%02x of %u bytes\n",
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vlr_subscr_name(conn->vsub),
GSM48_IE_AUTS, auts_tag, auts_len);
ran_conn_close(conn, GSM_CAUSE_AUTH_FAILED);
return -EINVAL;
}
if (msgb_l3len(msg) < sizeof(*gh) + 1 + 2 + auts_len) {
LOGP(DMM, LOGL_INFO,
"%s: MM R99 AUTHENTICATION FAILURE:"
" invalid Synch Failure msg: message truncated (%u)\n",
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vlr_subscr_name(conn->vsub), msgb_l3len(msg));
ran_conn_close(conn, GSM_CAUSE_AUTH_FAILED);
return -EINVAL;
}
/* We have an AUTS IE with exactly 14 bytes of AUTS and the msgb is
* large enough. */
DEBUGP(DMM, "%s: MM R99 AUTHENTICATION SYNCH (AUTS = %s)\n",
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
vlr_subscr_name(conn->vsub), osmo_hexdump_nospc(auts, 14));
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
return vlr_subscr_rx_auth_fail(conn->vsub, auts);
}
static int gsm48_rx_mm_tmsi_reall_compl(struct ran_conn *conn)
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
{
DEBUGP(DMM, "TMSI Reallocation Completed. Subscriber: %s\n",
vlr_subscr_name(conn->vsub));
if (!conn->vsub) {
LOGP(DMM, LOGL_ERROR,
"Rx MM TMSI Reallocation Complete: invalid: no subscriber\n");
return -EINVAL;
}
return vlr_subscr_rx_tmsi_reall_compl(conn->vsub);
}
2009-02-03 12:59:45 +00:00
/* Receive a GSM 04.08 Mobility Management (MM) message */
static int gsm0408_rcv_mm(struct ran_conn *conn, struct msgb *msg)
{
struct gsm48_hdr *gh = msgb_l3(msg);
int rc = 0;
switch (gsm48_hdr_msg_type(gh)) {
case GSM48_MT_MM_LOC_UPD_REQUEST:
rc = mm_rx_loc_upd_req(conn, msg);
break;
case GSM48_MT_MM_ID_RESP:
rc = mm_rx_id_resp(conn, msg);
break;
case GSM48_MT_MM_CM_SERV_REQ:
rc = gsm48_rx_mm_serv_req(conn, msg);
2008-12-27 01:55:51 +00:00
break;
case GSM48_MT_MM_STATUS:
rc = gsm48_rx_mm_status(msg);
break;
case GSM48_MT_MM_TMSI_REALL_COMPL:
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
rc = gsm48_rx_mm_tmsi_reall_compl(conn);
break;
2009-02-22 21:14:55 +00:00
case GSM48_MT_MM_IMSI_DETACH_IND:
rc = gsm48_rx_mm_imsi_detach_ind(conn, msg);
2009-02-22 21:14:55 +00:00
break;
case GSM48_MT_MM_CM_REEST_REQ:
rc = gsm48_rx_cm_reest_req(conn, msg);
2009-02-22 21:14:55 +00:00
break;
case GSM48_MT_MM_AUTH_RESP:
rc = gsm48_rx_mm_auth_resp(conn, msg);
break;
case GSM48_MT_MM_AUTH_FAIL:
rc = gsm48_rx_mm_auth_fail(conn, msg);
break;
default:
LOGP(DMM, LOGL_NOTICE, "Unknown GSM 04.08 MM msg type 0x%02x\n",
gh->msg_type);
break;
}
return rc;
}
2009-02-03 12:59:45 +00:00
2009-02-06 12:02:13 +00:00
/* Receive a PAGING RESPONSE message from the MS */
static int gsm48_rx_rr_pag_resp(struct ran_conn *conn, struct msgb *msg)
2009-02-06 12:02:13 +00:00
{
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
struct gsm_network *net = conn->network;
2009-02-06 12:02:13 +00:00
struct gsm48_hdr *gh = msgb_l3(msg);
store classmark in vlr_subscr, not conn Store all Classmark information in the VLR. So, we now always know the Classmark 1 (mandatory IE for LU). This is visible in the msc_vlr_tests -- they no longer indicate "assuming A5/1 is supported" because classmark 1 is missing, because we now know the Classmark 1. Rationale: During Location Updating, we receive Classmark 1; during CM Service Request and Paging Response, we receive Classmark 2. So far we stored these only for the duration of the conn, so as soon as a LU is complete, we would forget CM1. In other words, for anything else than a LU Request, we had no Classmark 1 available at all. During Ciphering Mode Command, we rely on Classmark 1 to determine whether A5/1 is supported. That is moot if we don't even have a Classmark 1 for any CM Service Request or Paging Response initiated connections. The only reason that A5/1 worked is that we assume A5/1 to work if Classmark 1 is missing. To add to the confusion, if a phone indicated that it did *not* support A5/1 in the Classmark 1, according to spec we're supposed to not service it at all. A code comment however says that we instead want to heed the flag -- which so far was only present in a Location Updating initiated connection. Now we can make this decision without assuming things. This got my attention while hacking on sending a BSSMAP Classmark Request from the MSC if it finds missing Classmark information, and was surprised to see it it lacking CM1 to decide about A5/1. Change-Id: I27081bf6e9e017923b2d02607f7ea06beddad82a
2018-09-13 01:05:52 +00:00
uint8_t classmark2_len = gh->data[1];
uint8_t *classmark2 = gh->data+2;
uint8_t *mi_lv = classmark2 + classmark2_len;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
struct osmo_location_area_id lai;
bool is_utran;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
lai.plmn = conn->network->plmn;
mscsplit: various preparations to separate MSC from BSC Disable large parts of the code that depend on BSC presence. The code sections disabled by #if BEFORE_MSCSPLIT shall be modified or dropped in the course of adding the A-interface. Don't set msg->lchan nor msg->dst. Don't use lchan in libmsc. Decouple lac from bts. Prepare entry/exit point for MSC -> BSC and MSC -> RNC communication: Add msc_ifaces.[hc], a_iface.c, with a general msc_tx_dtap() to redirect to different interfaces depending on the actual subscriber connection. While iu_tx() is going to be functional fairly soon, the a_tx() is going to be just a dummy for some time (see comment). Add Iu specific fields in gsm_subscriber_connection: the UE connection pointer and an indicator for the Integrity Protection status on Iu (to be fully implemented in later commits). Add lac member to gsm_subscriber_connection, to allow decoupling from bts->location_area_code. The conn->lac will actually be set in iu.c in an upcoming commit ("add iucs.[hc]"). move to libcommon-cs: gsm48_extract_mi(), gsm48_paging_extract_mi(). libmsc: duplicate gsm0808 / gsm48 functions (towards BSC). In osmo-nitb, libmsc would directly call the functions on the BSC level, not always via the bsc_api. When separating libmsc from libbsc, some functions are missing from the linkage. Hence duplicate these functions to libmsc, add an msc_ prefix for clarity, also add a _tx to gsm0808_cipher_mode(): * add msc_gsm0808_tx_cipher_mode() (dummy/stub) * add msc_gsm48_tx_mm_serv_ack() * add msc_gsm48_tx_mm_serv_rej() Call these from libmsc instead of * gsm0808_cipher_mode() * gsm48_tx_mm_serv_ack() * gsm48_tx_mm_serv_rej() Also add a comment related to msc_gsm0808_tx_cipher_mode() in two places. Remove internal RTP streaming code; OsmoNITB supported that, but for OsmoMSC, this will be done with an external MGCP gateway. Remove LCHAN_MODIFY from internal MNCC state machine. Temporarily disable all paging to be able to link libmsc without libbsc. Skip the paging part of channel_test because the paging is now disabled. Employ fake paging shims in order for msc_vlr_tests to still work. msc_compl_l3(): publish in .h, tweak return value. Use new libmsc enum values for return val, to avoid dependency on libbsc headers. Make callable from other scopes: publish in osmo_msc.h and remove 'static' in osmo_msc.c add gsm_encr to subscr_conn move subscr_request to gsm_subscriber.h subscr_request_channel() -> subscr_request_conn() move to libmsc: osmo_stats_vty_add_cmds() gsm_04_08: remove apply_codec_restrictions() gsm0408_test: use NULL for root ctx move to libbsc: gsm_bts_neighbor() move to libbsc: lchan_next_meas_rep() move vty config for t3212 to network level (periodic lu) remove unneccessary linking from some tests remove handle_abisip_signal() abis_rsl.c: don't use libvlr from libbsc gsm_subscriber_connection: put the LAC here, so that it is available without accessing conn->bts. In bsc_api.c, place this lac in conn for the sake of transition: Iu and A will use this new field to pass the LAC around, but in a completely separate OsmoBSC this is not actually needed. It can be removed again from osmo-bsc.git when the time has come. Siemens MRPCI: completely drop sending the MRPCI messages for now, they shall be added in osmo-bsc once the A-Interface code has settled. See OS#2389. Related: OS#1845 OS#2257 OS#2389 Change-Id: Id3705236350d5f69e447046b0a764bbabc3d493c
2017-05-08 13:12:20 +00:00
lai.lac = conn->lac;
2009-02-06 12:02:13 +00:00
if (ran_conn_is_establishing_auth_ciph(conn)) {
refactor subscr_conn and subscr_conn_fsm de-/alloc Refactor: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. 2. Add separate AUTH_CIPH state to the FSM. 3. Use conn->use_count to trigger conn release. 4. Add separate RELEASING state to the FSM. 5. Add rate counters for each of the three Complete Layer 3 types. Details: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. Historically, a gsm_subscriber_connection was allocated in libbsc land, and only upon Complete Layer 3 did libmsc add the fsm instance. After splitting openbsc.git into a separate osmo-msc, this is no longer necessary, hence: Closely tie gsm_subscriber_connection allocation to the subscr_conn_fsm instance: talloc the conn as a child of the FSM instance, and discard the conn as soon as the FSM terminates. 2. Add separate AUTH_CIPH state to the FSM. Decoding the Complete Layer 3 message is distinctly separate from waiting for the VLR FSMs to conclude. Use the NEW state as "we don't know if this is a valid message yet", and the AUTH_CIPH state as "evaluating, don't release". A profound effect of this: should we for any odd reason fail to leave the FSM's NEW state, the conn will be released right at the end of msc_compl_l3(), without needing to trigger release in each code path. 3. Use conn->use_count to trigger conn release. Before, the FSM itself would hold a use count on the conn, and hence we would need to ask it whether it is ready to release the conn yet by dispatching events, to achieve a use_count decrement. Instead, unite the FSM instance and conn, and do not hold a use count by the FSM. Hence, trigger an FSM "UNUSED" event only when the use_count reaches zero. As long as use counts are done correctly, the FSM will terminate correctly. These exceptions: - The new AUTH_CIPH state explicitly ignores UNUSED events, since we expect the use count to reach zero while evaluating Authentication and Ciphering. (I experimented with holding a use count by AUTH_CIPH onenter() and releasing by onleave(), but the use count and thus the conn are released before the next state can initiate transactions that would increment the use count again. Same thing for the VLR FSMs holding a use count, they should be done before we advance to the next state. The easiest is to simply expect zero use count during the AUTH_CIPH state.) - A CM Service Request means that even though the MSC would be through with all it wants to do, we shall still wait for a request to follow from the MS. Hence the FSM holds a use count on itself while a CM Service is pending. - While waiting for a Release/Clear Complete, the FSM holds a use count on itself. 4. Add separate RELEASING state to the FSM. If we decide to release for other reasons than a use count reaching zero, we still need to be able to wait for the msc_dtap() use count on the conn to release. (An upcoming patch will further use the RELEASING state to properly wait for Clear Complete / Release Complete messages.) 5. Add rate counters for each of the three Complete Layer 3 types. Besides LU, also count CM Service Request and Paging Response acceptance/rejections. Without these counters, only very few of the auth+ciph outcomes actually show in the counters. Related: OS#3122 Change-Id: I55feb379e176a96a831e105b86202b17a0ffe889
2018-03-30 22:02:14 +00:00
LOGP(DMM, LOGL_ERROR,
"Ignoring Paging Response, conn already busy establishing authenticity\n");
return 0;
}
if (ran_conn_is_accepted(conn)) {
refactor subscr_conn and subscr_conn_fsm de-/alloc Refactor: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. 2. Add separate AUTH_CIPH state to the FSM. 3. Use conn->use_count to trigger conn release. 4. Add separate RELEASING state to the FSM. 5. Add rate counters for each of the three Complete Layer 3 types. Details: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. Historically, a gsm_subscriber_connection was allocated in libbsc land, and only upon Complete Layer 3 did libmsc add the fsm instance. After splitting openbsc.git into a separate osmo-msc, this is no longer necessary, hence: Closely tie gsm_subscriber_connection allocation to the subscr_conn_fsm instance: talloc the conn as a child of the FSM instance, and discard the conn as soon as the FSM terminates. 2. Add separate AUTH_CIPH state to the FSM. Decoding the Complete Layer 3 message is distinctly separate from waiting for the VLR FSMs to conclude. Use the NEW state as "we don't know if this is a valid message yet", and the AUTH_CIPH state as "evaluating, don't release". A profound effect of this: should we for any odd reason fail to leave the FSM's NEW state, the conn will be released right at the end of msc_compl_l3(), without needing to trigger release in each code path. 3. Use conn->use_count to trigger conn release. Before, the FSM itself would hold a use count on the conn, and hence we would need to ask it whether it is ready to release the conn yet by dispatching events, to achieve a use_count decrement. Instead, unite the FSM instance and conn, and do not hold a use count by the FSM. Hence, trigger an FSM "UNUSED" event only when the use_count reaches zero. As long as use counts are done correctly, the FSM will terminate correctly. These exceptions: - The new AUTH_CIPH state explicitly ignores UNUSED events, since we expect the use count to reach zero while evaluating Authentication and Ciphering. (I experimented with holding a use count by AUTH_CIPH onenter() and releasing by onleave(), but the use count and thus the conn are released before the next state can initiate transactions that would increment the use count again. Same thing for the VLR FSMs holding a use count, they should be done before we advance to the next state. The easiest is to simply expect zero use count during the AUTH_CIPH state.) - A CM Service Request means that even though the MSC would be through with all it wants to do, we shall still wait for a request to follow from the MS. Hence the FSM holds a use count on itself while a CM Service is pending. - While waiting for a Release/Clear Complete, the FSM holds a use count on itself. 4. Add separate RELEASING state to the FSM. If we decide to release for other reasons than a use count reaching zero, we still need to be able to wait for the msc_dtap() use count on the conn to release. (An upcoming patch will further use the RELEASING state to properly wait for Clear Complete / Release Complete messages.) 5. Add rate counters for each of the three Complete Layer 3 types. Besides LU, also count CM Service Request and Paging Response acceptance/rejections. Without these counters, only very few of the auth+ciph outcomes actually show in the counters. Related: OS#3122 Change-Id: I55feb379e176a96a831e105b86202b17a0ffe889
2018-03-30 22:02:14 +00:00
LOGP(DMM, LOGL_ERROR, "Ignoring Paging Response, conn already established\n");
return 0;
}
refactor log ctx for vlr_subscr and ran_conn ran_conn_get_conn_id(): instead of a talloc allocated string, return a static buffer in ran_conn_get_conn_id(). So far this function had no callers. Refactor ran_conn_update_id() API: during early L3-Complete, when no subscriber is associated yet, update the FSM Id by the MI type seen in the L3 Complete message: ran_conn_update_id_from_mi(). Later on set the vsub and re-update. Call vlr.ops->subscr_update when the TMSI is updated, so that log context includes the TMSI from then on. Enrich context for vlr_subscr_name and ran_conn fi name. Include all available information in vlr_subscr_name(); instead of either IMSI or MSISDN or TMSI, print all of them when present. Instead of a short log, rather have more valuable context. A context info would now look like: Process_Access_Request_VLR(IMSI-901700000014706:MSISDN-2023:TMSI-0x08BDE4EC:GERAN-A-3:PAGING_RESP) It does get quite long, but ensures easy correlation of any BSSAP / IuCS messages with log output, especially if multiple subscribers are busy at the same time. Print TMSI and TMSInew in uppercase hexadecimal, which is the typical representation in the telecom world. When showing the RAN conn id GERAN_A-00000017 becomes GERAN-A-23 - We usually write the conn_id in decimal. - Leading zeros are clutter and might suggest hexadecimal format. - 'GERAN-A' and 'UTRAN-Iu' are the strings defined by osmo_rat_type_name(). Depends: I7798c3ef983c2e333b2b9cbffef6f366f370bd81 (libosmocore) Depends: Ica25919758ef6cba8348da199b0ae7e0ba628798 (libosmocore) Change-Id: I66a68ce2eb8957a35855a3743d91a86299900834
2019-01-03 01:32:14 +00:00
conn->complete_layer3_type = COMPLETE_LAYER3_PAGING_RESP;
ran_conn_update_id_from_mi(conn, mi_lv + 1, *mi_lv);
LOG_RAN_CONN_CAT(conn, DRR, LOGL_DEBUG, "PAGING RESPONSE\n");
is_utran = (conn->via_ran == OSMO_RAT_UTRAN_IU);
vlr_proc_acc_req(conn->fi,
RAN_CONN_E_ACCEPTED, RAN_CONN_E_CN_CLOSE, NULL,
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
net->vlr, conn,
VLR_PR_ARQ_T_PAGING_RESP, mi_lv, &lai,
is_utran || conn->network->authentication_required,
is_utran || conn->network->a5_encryption_mask > 0x01,
store classmark in vlr_subscr, not conn Store all Classmark information in the VLR. So, we now always know the Classmark 1 (mandatory IE for LU). This is visible in the msc_vlr_tests -- they no longer indicate "assuming A5/1 is supported" because classmark 1 is missing, because we now know the Classmark 1. Rationale: During Location Updating, we receive Classmark 1; during CM Service Request and Paging Response, we receive Classmark 2. So far we stored these only for the duration of the conn, so as soon as a LU is complete, we would forget CM1. In other words, for anything else than a LU Request, we had no Classmark 1 available at all. During Ciphering Mode Command, we rely on Classmark 1 to determine whether A5/1 is supported. That is moot if we don't even have a Classmark 1 for any CM Service Request or Paging Response initiated connections. The only reason that A5/1 worked is that we assume A5/1 to work if Classmark 1 is missing. To add to the confusion, if a phone indicated that it did *not* support A5/1 in the Classmark 1, according to spec we're supposed to not service it at all. A code comment however says that we instead want to heed the flag -- which so far was only present in a Location Updating initiated connection. Now we can make this decision without assuming things. This got my attention while hacking on sending a BSSMAP Classmark Request from the MSC if it finds missing Classmark information, and was surprised to see it it lacking CM1 to decide about A5/1. Change-Id: I27081bf6e9e017923b2d02607f7ea06beddad82a
2018-09-13 01:05:52 +00:00
classmark2_is_r99(classmark2, classmark2_len),
is_utran);
store classmark in vlr_subscr, not conn Store all Classmark information in the VLR. So, we now always know the Classmark 1 (mandatory IE for LU). This is visible in the msc_vlr_tests -- they no longer indicate "assuming A5/1 is supported" because classmark 1 is missing, because we now know the Classmark 1. Rationale: During Location Updating, we receive Classmark 1; during CM Service Request and Paging Response, we receive Classmark 2. So far we stored these only for the duration of the conn, so as soon as a LU is complete, we would forget CM1. In other words, for anything else than a LU Request, we had no Classmark 1 available at all. During Ciphering Mode Command, we rely on Classmark 1 to determine whether A5/1 is supported. That is moot if we don't even have a Classmark 1 for any CM Service Request or Paging Response initiated connections. The only reason that A5/1 worked is that we assume A5/1 to work if Classmark 1 is missing. To add to the confusion, if a phone indicated that it did *not* support A5/1 in the Classmark 1, according to spec we're supposed to not service it at all. A code comment however says that we instead want to heed the flag -- which so far was only present in a Location Updating initiated connection. Now we can make this decision without assuming things. This got my attention while hacking on sending a BSSMAP Classmark Request from the MSC if it finds missing Classmark information, and was surprised to see it it lacking CM1 to decide about A5/1. Change-Id: I27081bf6e9e017923b2d02607f7ea06beddad82a
2018-09-13 01:05:52 +00:00
/* From vlr_proc_acc_req() we expect an implicit dispatch of PR_ARQ_E_START we expect
* msc_vlr_subscr_assoc() to already have been called and completed. Has an error occured? */
if (!conn->vsub) {
LOG_RAN_CONN(conn, LOGL_ERROR, "subscriber not allowed to do a Paging Response\n");
store classmark in vlr_subscr, not conn Store all Classmark information in the VLR. So, we now always know the Classmark 1 (mandatory IE for LU). This is visible in the msc_vlr_tests -- they no longer indicate "assuming A5/1 is supported" because classmark 1 is missing, because we now know the Classmark 1. Rationale: During Location Updating, we receive Classmark 1; during CM Service Request and Paging Response, we receive Classmark 2. So far we stored these only for the duration of the conn, so as soon as a LU is complete, we would forget CM1. In other words, for anything else than a LU Request, we had no Classmark 1 available at all. During Ciphering Mode Command, we rely on Classmark 1 to determine whether A5/1 is supported. That is moot if we don't even have a Classmark 1 for any CM Service Request or Paging Response initiated connections. The only reason that A5/1 worked is that we assume A5/1 to work if Classmark 1 is missing. To add to the confusion, if a phone indicated that it did *not* support A5/1 in the Classmark 1, according to spec we're supposed to not service it at all. A code comment however says that we instead want to heed the flag -- which so far was only present in a Location Updating initiated connection. Now we can make this decision without assuming things. This got my attention while hacking on sending a BSSMAP Classmark Request from the MSC if it finds missing Classmark information, and was surprised to see it it lacking CM1 to decide about A5/1. Change-Id: I27081bf6e9e017923b2d02607f7ea06beddad82a
2018-09-13 01:05:52 +00:00
return -EIO;
}
memcpy(conn->vsub->classmark.classmark2, classmark2, classmark2_len);
conn->vsub->classmark.classmark2_len = classmark2_len;
ran_conn_complete_layer_3(conn);
return 0;
2009-02-06 12:02:13 +00:00
}
static int gsm48_rx_rr_app_info(struct ran_conn *conn, struct msgb *msg)
{
struct gsm48_hdr *gh = msgb_l3(msg);
uint8_t apdu_id_flags;
uint8_t apdu_len;
uint8_t *apdu_data;
apdu_id_flags = gh->data[0];
apdu_len = gh->data[1];
apdu_data = gh->data+2;
DEBUGP(DRR, "RX APPLICATION INFO id/flags=0x%02x apdu_len=%u apdu=%s\n",
apdu_id_flags, apdu_len, osmo_hexdump(apdu_data, apdu_len));
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
/* we're not using the app info blob anywhere, so ignore. */
#if 0
return db_apdu_blob_store(conn->subscr, apdu_id_flags, apdu_len, apdu_data);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
#else
return 0;
#endif
}
2009-02-03 12:59:45 +00:00
/* Receive a GSM 04.08 Radio Resource (RR) message */
static int gsm0408_rcv_rr(struct ran_conn *conn, struct msgb *msg)
{
struct gsm48_hdr *gh = msgb_l3(msg);
2009-02-06 12:02:13 +00:00
int rc = 0;
switch (gh->msg_type) {
case GSM48_MT_RR_PAG_RESP:
rc = gsm48_rx_rr_pag_resp(conn, msg);
2009-02-06 12:02:13 +00:00
break;
case GSM48_MT_RR_APP_INFO:
rc = gsm48_rx_rr_app_info(conn, msg);
break;
default:
LOGP(DRR, LOGL_NOTICE, "MSC: Unimplemented %s GSM 04.08 RR "
"message\n", gsm48_rr_msg_name(gh->msg_type));
break;
}
2009-02-06 12:02:13 +00:00
return rc;
}
int gsm48_send_rr_app_info(struct ran_conn *conn, uint8_t apdu_id,
uint8_t apdu_len, const uint8_t *apdu)
{
struct msgb *msg = gsm48_msgb_alloc_name("GSM 04.08 APP INF");
struct gsm48_hdr *gh;
DEBUGP(DRR, "TX APPLICATION INFO id=0x%02x, len=%u\n",
apdu_id, apdu_len);
gh = (struct gsm48_hdr *) msgb_put(msg, sizeof(*gh) + 2 + apdu_len);
gh->proto_discr = GSM48_PDISC_RR;
gh->msg_type = GSM48_MT_RR_APP_INFO;
gh->data[0] = apdu_id;
gh->data[1] = apdu_len;
memcpy(gh->data+2, apdu, apdu_len);
return gsm48_conn_sendmsg(msg, conn, NULL);
}
static bool msg_is_initially_permitted(const struct gsm48_hdr *hdr)
{
uint8_t pdisc = gsm48_hdr_pdisc(hdr);
uint8_t msg_type = gsm48_hdr_msg_type(hdr);
switch (pdisc) {
case GSM48_PDISC_MM:
switch (msg_type) {
case GSM48_MT_MM_LOC_UPD_REQUEST:
case GSM48_MT_MM_CM_SERV_REQ:
case GSM48_MT_MM_CM_REEST_REQ:
case GSM48_MT_MM_AUTH_RESP:
case GSM48_MT_MM_AUTH_FAIL:
case GSM48_MT_MM_ID_RESP:
case GSM48_MT_MM_TMSI_REALL_COMPL:
case GSM48_MT_MM_IMSI_DETACH_IND:
return true;
default:
break;
}
break;
case GSM48_PDISC_RR:
switch (msg_type) {
/* GSM48_MT_RR_CIPH_M_COMPL is actually handled in bssmap_rx_ciph_compl() and gets redirected in the
* BSSAP layer to ran_conn_cipher_mode_compl() (before this here is reached) */
case GSM48_MT_RR_PAG_RESP:
return true;
default:
break;
}
break;
default:
break;
}
return false;
}
void cm_service_request_concludes(struct ran_conn *conn,
struct msgb *msg)
2008-12-27 16:32:52 +00:00
{
/* If a CM Service Request was received before, this is the request the
* conn was opened for. No need to wait for further messages. */
2008-12-27 16:32:52 +00:00
if (!conn->received_cm_service_request)
return;
2008-12-27 16:32:52 +00:00
if (log_check_level(DMM, LOGL_DEBUG)) {
struct gsm48_hdr *gh = msgb_l3(msg);
uint8_t pdisc = gsm48_hdr_pdisc(gh);
uint8_t msg_type = gsm48_hdr_msg_type(gh);
2008-12-27 16:32:52 +00:00
DEBUGP(DMM, "%s: rx msg %s:"
" received_cm_service_request changes to false\n",
vlr_subscr_name(conn->vsub),
gsm48_pdisc_msgtype_name(pdisc, msg_type));
}
conn->received_cm_service_request = false;
ran_conn_put(conn, RAN_CONN_USE_CM_SERVICE);
}
/* TS 24.007 11.2.3.2.3 Message Type Octet / Duplicate Detection */
int gsm0407_pdisc_ctr_bin(uint8_t pdisc)
{
switch (pdisc) {
case GSM48_PDISC_MM:
case GSM48_PDISC_CC:
case GSM48_PDISC_NC_SS:
return 0;
case GSM48_PDISC_GROUP_CC:
return 1;
case GSM48_PDISC_BCAST_CC:
return 2;
case GSM48_PDISC_LOC:
return 3;
default:
return -1;
}
2008-12-27 16:32:52 +00:00
}
/* extract the N(SD) and return the modulo value for a R98 message */
static uint8_t gsm0407_determine_nsd_ret_modulo_r99(uint8_t pdisc, uint8_t msg_type, uint8_t *n_sd)
{
switch (pdisc) {
case GSM48_PDISC_MM:
case GSM48_PDISC_CC:
case GSM48_PDISC_NC_SS:
*n_sd = (msg_type >> 6) & 0x3;
return 4;
case GSM48_PDISC_GROUP_CC:
case GSM48_PDISC_BCAST_CC:
case GSM48_PDISC_LOC:
*n_sd = (msg_type >> 6) & 0x1;
return 2;
default:
/* no sequence number, we cannot detect dups */
return 0;
}
}
/* extract the N(SD) and return the modulo value for a R99 message */
static uint8_t gsm0407_determine_nsd_ret_modulo_r98(uint8_t pdisc, uint8_t msg_type, uint8_t *n_sd)
{
switch (pdisc) {
case GSM48_PDISC_MM:
case GSM48_PDISC_CC:
case GSM48_PDISC_NC_SS:
case GSM48_PDISC_GROUP_CC:
case GSM48_PDISC_BCAST_CC:
case GSM48_PDISC_LOC:
*n_sd = (msg_type >> 6) & 0x1;
return 2;
default:
/* no sequence number, we cannot detect dups */
return 0;
}
}
/* TS 24.007 11.2.3.2 Message Type Octet / Duplicate Detection */
static bool gsm0407_is_duplicate(struct ran_conn *conn, struct msgb *msg)
{
struct gsm48_hdr *gh;
uint8_t pdisc;
uint8_t n_sd, modulo;
int bin;
gh = msgb_l3(msg);
pdisc = gsm48_hdr_pdisc(gh);
store classmark in vlr_subscr, not conn Store all Classmark information in the VLR. So, we now always know the Classmark 1 (mandatory IE for LU). This is visible in the msc_vlr_tests -- they no longer indicate "assuming A5/1 is supported" because classmark 1 is missing, because we now know the Classmark 1. Rationale: During Location Updating, we receive Classmark 1; during CM Service Request and Paging Response, we receive Classmark 2. So far we stored these only for the duration of the conn, so as soon as a LU is complete, we would forget CM1. In other words, for anything else than a LU Request, we had no Classmark 1 available at all. During Ciphering Mode Command, we rely on Classmark 1 to determine whether A5/1 is supported. That is moot if we don't even have a Classmark 1 for any CM Service Request or Paging Response initiated connections. The only reason that A5/1 worked is that we assume A5/1 to work if Classmark 1 is missing. To add to the confusion, if a phone indicated that it did *not* support A5/1 in the Classmark 1, according to spec we're supposed to not service it at all. A code comment however says that we instead want to heed the flag -- which so far was only present in a Location Updating initiated connection. Now we can make this decision without assuming things. This got my attention while hacking on sending a BSSMAP Classmark Request from the MSC if it finds missing Classmark information, and was surprised to see it it lacking CM1 to decide about A5/1. Change-Id: I27081bf6e9e017923b2d02607f7ea06beddad82a
2018-09-13 01:05:52 +00:00
if (conn->vsub && classmark_is_r99(&conn->vsub->classmark)) {
modulo = gsm0407_determine_nsd_ret_modulo_r99(pdisc, gh->msg_type, &n_sd);
store classmark in vlr_subscr, not conn Store all Classmark information in the VLR. So, we now always know the Classmark 1 (mandatory IE for LU). This is visible in the msc_vlr_tests -- they no longer indicate "assuming A5/1 is supported" because classmark 1 is missing, because we now know the Classmark 1. Rationale: During Location Updating, we receive Classmark 1; during CM Service Request and Paging Response, we receive Classmark 2. So far we stored these only for the duration of the conn, so as soon as a LU is complete, we would forget CM1. In other words, for anything else than a LU Request, we had no Classmark 1 available at all. During Ciphering Mode Command, we rely on Classmark 1 to determine whether A5/1 is supported. That is moot if we don't even have a Classmark 1 for any CM Service Request or Paging Response initiated connections. The only reason that A5/1 worked is that we assume A5/1 to work if Classmark 1 is missing. To add to the confusion, if a phone indicated that it did *not* support A5/1 in the Classmark 1, according to spec we're supposed to not service it at all. A code comment however says that we instead want to heed the flag -- which so far was only present in a Location Updating initiated connection. Now we can make this decision without assuming things. This got my attention while hacking on sending a BSSMAP Classmark Request from the MSC if it finds missing Classmark information, and was surprised to see it it lacking CM1 to decide about A5/1. Change-Id: I27081bf6e9e017923b2d02607f7ea06beddad82a
2018-09-13 01:05:52 +00:00
} else { /* pre R99 */
modulo = gsm0407_determine_nsd_ret_modulo_r98(pdisc, gh->msg_type, &n_sd);
}
if (modulo == 0)
return false;
bin = gsm0407_pdisc_ctr_bin(pdisc);
if (bin < 0)
return false;
OSMO_ASSERT(bin < ARRAY_SIZE(conn->n_sd_next));
if (n_sd != conn->n_sd_next[bin]) {
/* not what we expected: duplicate */
return true;
} else {
/* as expected: no dup; update expected counter for next message */
conn->n_sd_next[bin] = (n_sd + 1) % modulo;
return false;
2008-12-27 16:32:52 +00:00
}
}
extern int gsm0408_rcv_cc(struct ran_conn *conn, struct msgb *msg);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
/* Main entry point for GSM 04.08/44.008 Layer 3 data (e.g. from the BSC). */
int gsm0408_dispatch(struct ran_conn *conn, struct msgb *msg)
{
struct gsm48_hdr *gh;
uint8_t pdisc;
int rc = 0;
OSMO_ASSERT(msg->l3h);
OSMO_ASSERT(conn);
OSMO_ASSERT(msg);
gh = msgb_l3(msg);
pdisc = gsm48_hdr_pdisc(gh);
if (gsm0407_is_duplicate(conn, msg)) {
LOGP(DRLL, LOGL_NOTICE, "%s: Discarding duplicate L3 message\n",
(conn && conn->vsub) ? vlr_subscr_name(conn->vsub) : "UNKNOWN");
return 0;
}
LOGP(DRLL, LOGL_DEBUG, "Dispatching 04.08 message %s (0x%x:0x%x)\n",
gsm48_pdisc_msgtype_name(pdisc, gsm48_hdr_msg_type(gh)),
pdisc, gsm48_hdr_msg_type(gh));
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
if (!ran_conn_is_accepted(conn)
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
&& !msg_is_initially_permitted(gh)) {
LOGP(DRLL, LOGL_ERROR,
"subscr %s: Message not permitted for initial conn: %s\n",
vlr_subscr_name(conn->vsub),
gsm48_pdisc_msgtype_name(pdisc, gsm48_hdr_msg_type(gh)));
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
return -EACCES;
}
if (conn->vsub && conn->vsub->cs.attached_via_ran != conn->via_ran) {
LOGP(DMM, LOGL_ERROR,
"%s: Illegal situation: RAN type mismatch:"
" attached via %s, received message via %s\n",
vlr_subscr_name(conn->vsub),
osmo_rat_type_name(conn->vsub->cs.attached_via_ran),
osmo_rat_type_name(conn->via_ran));
return -EACCES;
}
#if 0
if (silent_call_reroute(conn, msg))
return silent_call_rx(conn, msg);
#endif
switch (pdisc) {
case GSM48_PDISC_CC:
rc = gsm0408_rcv_cc(conn, msg);
break;
case GSM48_PDISC_MM:
rc = gsm0408_rcv_mm(conn, msg);
break;
case GSM48_PDISC_RR:
rc = gsm0408_rcv_rr(conn, msg);
break;
case GSM48_PDISC_SMS:
rc = gsm0411_rcv_sms(conn, msg);
break;
case GSM48_PDISC_MM_GPRS:
case GSM48_PDISC_SM_GPRS:
LOGP(DRLL, LOGL_NOTICE, "Unimplemented "
"GSM 04.08 discriminator 0x%02x\n", pdisc);
rc = -ENOTSUP;
break;
case GSM48_PDISC_NC_SS:
rc = gsm0911_rcv_nc_ss(conn, msg);
break;
case GSM48_PDISC_TEST:
rc = gsm0414_rcv_test(conn, msg);
break;
default:
LOGP(DRLL, LOGL_NOTICE, "Unknown "
"GSM 04.08 discriminator 0x%02x\n", pdisc);
rc = -EINVAL;
break;
}
return rc;
}
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
/***********************************************************************
* VLR integration
***********************************************************************/
/* VLR asks us to send an authentication request */
static int msc_vlr_tx_auth_req(void *msc_conn_ref, struct vlr_auth_tuple *at,
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
bool send_autn)
{
struct ran_conn *conn = msc_conn_ref;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
return gsm48_tx_mm_auth_req(conn, at->vec.rand,
send_autn? at->vec.autn : NULL,
at->key_seq);
}
/* VLR asks us to send an authentication reject */
static int msc_vlr_tx_auth_rej(void *msc_conn_ref)
{
struct ran_conn *conn = msc_conn_ref;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
return gsm48_tx_mm_auth_rej(conn);
}
/* VLR asks us to transmit an Identity Request of given type */
static int msc_vlr_tx_id_req(void *msc_conn_ref, uint8_t mi_type)
{
struct ran_conn *conn = msc_conn_ref;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
return mm_tx_identity_req(conn, mi_type);
}
/* VLR asks us to transmit a Location Update Accept */
static int msc_vlr_tx_lu_acc(void *msc_conn_ref, uint32_t send_tmsi)
{
struct ran_conn *conn = msc_conn_ref;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
return gsm0408_loc_upd_acc(conn, send_tmsi);
}
/* VLR asks us to transmit a Location Update Reject */
static int msc_vlr_tx_lu_rej(void *msc_conn_ref, enum gsm48_reject_value cause)
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
{
struct ran_conn *conn = msc_conn_ref;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
return gsm0408_loc_upd_rej(conn, cause);
}
/* VLR asks us to transmit a CM Service Accept */
static int msc_vlr_tx_cm_serv_acc(void *msc_conn_ref)
{
struct ran_conn *conn = msc_conn_ref;
return msc_gsm48_tx_mm_serv_ack(conn);
}
static int msc_vlr_tx_common_id(void *msc_conn_ref)
{
struct ran_conn *conn = msc_conn_ref;
return msc_tx_common_id(conn);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
}
/* VLR asks us to transmit MM info. */
static int msc_vlr_tx_mm_info(void *msc_conn_ref)
{
struct ran_conn *conn = msc_conn_ref;
if (!conn->network->send_mm_info)
return 0;
return gsm48_tx_mm_info(conn);
}
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
/* VLR asks us to transmit a CM Service Reject */
static int msc_vlr_tx_cm_serv_rej(void *msc_conn_ref, enum gsm48_reject_value cause)
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
{
struct ran_conn *conn = msc_conn_ref;
refactor subscr_conn and subscr_conn_fsm de-/alloc Refactor: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. 2. Add separate AUTH_CIPH state to the FSM. 3. Use conn->use_count to trigger conn release. 4. Add separate RELEASING state to the FSM. 5. Add rate counters for each of the three Complete Layer 3 types. Details: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. Historically, a gsm_subscriber_connection was allocated in libbsc land, and only upon Complete Layer 3 did libmsc add the fsm instance. After splitting openbsc.git into a separate osmo-msc, this is no longer necessary, hence: Closely tie gsm_subscriber_connection allocation to the subscr_conn_fsm instance: talloc the conn as a child of the FSM instance, and discard the conn as soon as the FSM terminates. 2. Add separate AUTH_CIPH state to the FSM. Decoding the Complete Layer 3 message is distinctly separate from waiting for the VLR FSMs to conclude. Use the NEW state as "we don't know if this is a valid message yet", and the AUTH_CIPH state as "evaluating, don't release". A profound effect of this: should we for any odd reason fail to leave the FSM's NEW state, the conn will be released right at the end of msc_compl_l3(), without needing to trigger release in each code path. 3. Use conn->use_count to trigger conn release. Before, the FSM itself would hold a use count on the conn, and hence we would need to ask it whether it is ready to release the conn yet by dispatching events, to achieve a use_count decrement. Instead, unite the FSM instance and conn, and do not hold a use count by the FSM. Hence, trigger an FSM "UNUSED" event only when the use_count reaches zero. As long as use counts are done correctly, the FSM will terminate correctly. These exceptions: - The new AUTH_CIPH state explicitly ignores UNUSED events, since we expect the use count to reach zero while evaluating Authentication and Ciphering. (I experimented with holding a use count by AUTH_CIPH onenter() and releasing by onleave(), but the use count and thus the conn are released before the next state can initiate transactions that would increment the use count again. Same thing for the VLR FSMs holding a use count, they should be done before we advance to the next state. The easiest is to simply expect zero use count during the AUTH_CIPH state.) - A CM Service Request means that even though the MSC would be through with all it wants to do, we shall still wait for a request to follow from the MS. Hence the FSM holds a use count on itself while a CM Service is pending. - While waiting for a Release/Clear Complete, the FSM holds a use count on itself. 4. Add separate RELEASING state to the FSM. If we decide to release for other reasons than a use count reaching zero, we still need to be able to wait for the msc_dtap() use count on the conn to release. (An upcoming patch will further use the RELEASING state to properly wait for Clear Complete / Release Complete messages.) 5. Add rate counters for each of the three Complete Layer 3 types. Besides LU, also count CM Service Request and Paging Response acceptance/rejections. Without these counters, only very few of the auth+ciph outcomes actually show in the counters. Related: OS#3122 Change-Id: I55feb379e176a96a831e105b86202b17a0ffe889
2018-03-30 22:02:14 +00:00
int rc;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
refactor subscr_conn and subscr_conn_fsm de-/alloc Refactor: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. 2. Add separate AUTH_CIPH state to the FSM. 3. Use conn->use_count to trigger conn release. 4. Add separate RELEASING state to the FSM. 5. Add rate counters for each of the three Complete Layer 3 types. Details: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. Historically, a gsm_subscriber_connection was allocated in libbsc land, and only upon Complete Layer 3 did libmsc add the fsm instance. After splitting openbsc.git into a separate osmo-msc, this is no longer necessary, hence: Closely tie gsm_subscriber_connection allocation to the subscr_conn_fsm instance: talloc the conn as a child of the FSM instance, and discard the conn as soon as the FSM terminates. 2. Add separate AUTH_CIPH state to the FSM. Decoding the Complete Layer 3 message is distinctly separate from waiting for the VLR FSMs to conclude. Use the NEW state as "we don't know if this is a valid message yet", and the AUTH_CIPH state as "evaluating, don't release". A profound effect of this: should we for any odd reason fail to leave the FSM's NEW state, the conn will be released right at the end of msc_compl_l3(), without needing to trigger release in each code path. 3. Use conn->use_count to trigger conn release. Before, the FSM itself would hold a use count on the conn, and hence we would need to ask it whether it is ready to release the conn yet by dispatching events, to achieve a use_count decrement. Instead, unite the FSM instance and conn, and do not hold a use count by the FSM. Hence, trigger an FSM "UNUSED" event only when the use_count reaches zero. As long as use counts are done correctly, the FSM will terminate correctly. These exceptions: - The new AUTH_CIPH state explicitly ignores UNUSED events, since we expect the use count to reach zero while evaluating Authentication and Ciphering. (I experimented with holding a use count by AUTH_CIPH onenter() and releasing by onleave(), but the use count and thus the conn are released before the next state can initiate transactions that would increment the use count again. Same thing for the VLR FSMs holding a use count, they should be done before we advance to the next state. The easiest is to simply expect zero use count during the AUTH_CIPH state.) - A CM Service Request means that even though the MSC would be through with all it wants to do, we shall still wait for a request to follow from the MS. Hence the FSM holds a use count on itself while a CM Service is pending. - While waiting for a Release/Clear Complete, the FSM holds a use count on itself. 4. Add separate RELEASING state to the FSM. If we decide to release for other reasons than a use count reaching zero, we still need to be able to wait for the msc_dtap() use count on the conn to release. (An upcoming patch will further use the RELEASING state to properly wait for Clear Complete / Release Complete messages.) 5. Add rate counters for each of the three Complete Layer 3 types. Besides LU, also count CM Service Request and Paging Response acceptance/rejections. Without these counters, only very few of the auth+ciph outcomes actually show in the counters. Related: OS#3122 Change-Id: I55feb379e176a96a831e105b86202b17a0ffe889
2018-03-30 22:02:14 +00:00
rc = msc_gsm48_tx_mm_serv_rej(conn, cause);
if (conn->received_cm_service_request) {
conn->received_cm_service_request = false;
ran_conn_put(conn, RAN_CONN_USE_CM_SERVICE);
refactor subscr_conn and subscr_conn_fsm de-/alloc Refactor: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. 2. Add separate AUTH_CIPH state to the FSM. 3. Use conn->use_count to trigger conn release. 4. Add separate RELEASING state to the FSM. 5. Add rate counters for each of the three Complete Layer 3 types. Details: 1. Glue the gsm_subscriber_connection alloc to the subscr_conn_fsm. Historically, a gsm_subscriber_connection was allocated in libbsc land, and only upon Complete Layer 3 did libmsc add the fsm instance. After splitting openbsc.git into a separate osmo-msc, this is no longer necessary, hence: Closely tie gsm_subscriber_connection allocation to the subscr_conn_fsm instance: talloc the conn as a child of the FSM instance, and discard the conn as soon as the FSM terminates. 2. Add separate AUTH_CIPH state to the FSM. Decoding the Complete Layer 3 message is distinctly separate from waiting for the VLR FSMs to conclude. Use the NEW state as "we don't know if this is a valid message yet", and the AUTH_CIPH state as "evaluating, don't release". A profound effect of this: should we for any odd reason fail to leave the FSM's NEW state, the conn will be released right at the end of msc_compl_l3(), without needing to trigger release in each code path. 3. Use conn->use_count to trigger conn release. Before, the FSM itself would hold a use count on the conn, and hence we would need to ask it whether it is ready to release the conn yet by dispatching events, to achieve a use_count decrement. Instead, unite the FSM instance and conn, and do not hold a use count by the FSM. Hence, trigger an FSM "UNUSED" event only when the use_count reaches zero. As long as use counts are done correctly, the FSM will terminate correctly. These exceptions: - The new AUTH_CIPH state explicitly ignores UNUSED events, since we expect the use count to reach zero while evaluating Authentication and Ciphering. (I experimented with holding a use count by AUTH_CIPH onenter() and releasing by onleave(), but the use count and thus the conn are released before the next state can initiate transactions that would increment the use count again. Same thing for the VLR FSMs holding a use count, they should be done before we advance to the next state. The easiest is to simply expect zero use count during the AUTH_CIPH state.) - A CM Service Request means that even though the MSC would be through with all it wants to do, we shall still wait for a request to follow from the MS. Hence the FSM holds a use count on itself while a CM Service is pending. - While waiting for a Release/Clear Complete, the FSM holds a use count on itself. 4. Add separate RELEASING state to the FSM. If we decide to release for other reasons than a use count reaching zero, we still need to be able to wait for the msc_dtap() use count on the conn to release. (An upcoming patch will further use the RELEASING state to properly wait for Clear Complete / Release Complete messages.) 5. Add rate counters for each of the three Complete Layer 3 types. Besides LU, also count CM Service Request and Paging Response acceptance/rejections. Without these counters, only very few of the auth+ciph outcomes actually show in the counters. Related: OS#3122 Change-Id: I55feb379e176a96a831e105b86202b17a0ffe889
2018-03-30 22:02:14 +00:00
}
return rc;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
}
/* For msc_vlr_set_ciph_mode() */
osmo_static_assert(sizeof(((struct gsm0808_encrypt_info*)0)->key) >= sizeof(((struct osmo_auth_vector*)0)->kc),
gsm0808_encrypt_info_key_fits_osmo_auth_vec_kc);
int ran_conn_geran_set_cipher_mode(struct ran_conn *conn, bool umts_aka, bool retrieve_imeisv)
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
{
struct gsm_network *net;
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
struct gsm0808_encrypt_info ei;
int i, j = 0;
int request_classmark = 0;
int request_classmark_for_a5_n = 0;
struct vlr_auth_tuple *tuple;
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
if (!conn || !conn->vsub || !conn->vsub->last_tuple) {
/* This should really never happen, because we checked this in msc_vlr_set_ciph_mode()
* already. */
LOGP(DMM, LOGL_ERROR, "Internal error: missing state during Ciphering Mode Command\n");
return -EINVAL;
}
net = conn->network;
tuple = conn->vsub->last_tuple;
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
for (i = 0; i < 8; i++) {
int supported;
/* A5/n permitted by osmo-msc.cfg? */
if (!(net->a5_encryption_mask & (1 << i)))
continue;
/* A5/n supported by MS? */
supported = classmark_supports_a5(&conn->vsub->classmark, i);
if (supported == 1) {
ei.perm_algo[j++] = vlr_ciph_to_gsm0808_alg_id(i);
/* A higher A5/n is supported, so no need to request a Classmark
* for support of a lesser A5/n. */
request_classmark = 0;
} else if (supported < 0) {
request_classmark = -supported;
request_classmark_for_a5_n = i;
}
}
ei.perm_algo_len = j;
if (request_classmark) {
/* The highest A5/n as from osmo-msc.cfg might be available, but we are
* still missing the Classmark information for that from the MS. First
* ask for that. */
LOGP(DMM, LOGL_DEBUG, "%s: to determine whether A5/%d is supported,"
" first ask for a Classmark Update to obtain Classmark %d\n",
vlr_subscr_name(conn->vsub), request_classmark_for_a5_n,
request_classmark);
return ran_conn_classmark_request_then_cipher_mode_cmd(conn, umts_aka, retrieve_imeisv);
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
}
if (ei.perm_algo_len == 0) {
LOGP(DMM, LOGL_ERROR, "%s: cannot start ciphering, no intersection "
"between MSC-configured and MS-supported A5 algorithms. MSC: %x MS: %s\n",
vlr_subscr_name(conn->vsub), net->a5_encryption_mask,
classmark_a5_name(&conn->vsub->classmark));
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
return -ENOTSUP;
}
DEBUGP(DMM, "-> CIPHER MODE COMMAND %s\n", vlr_subscr_name(conn->vsub));
tuple = conn->vsub->last_tuple;
/* In case of UMTS AKA, the Kc for ciphering must be derived from the 3G auth
* tokens. tuple->vec.kc was calculated from the GSM algorithm and is not
* necessarily a match for the UMTS AKA tokens. */
if (umts_aka)
osmo_auth_c3(ei.key, tuple->vec.ck, tuple->vec.ik);
else
memcpy(ei.key, tuple->vec.kc, sizeof(tuple->vec.kc));
ei.key_len = sizeof(tuple->vec.kc);
conn->geran_encr = (struct geran_encr){};
if (ei.key_len <= sizeof(conn->geran_encr.key)) {
memcpy(conn->geran_encr.key, ei.key, ei.key_len);
conn->geran_encr.key_len = ei.key_len;
}
/* conn->geran_encr.alg_id remains unknown until we receive a Cipher Mode Complete from the BSC */
A5/n Ciph: request Classmark Update if missing When the VLR requests a Ciphering Mode with vlr_ops.set_ciph_mode(), and if we need a ciph algo flag from a Classmark information that is not yet known (usually CM 2 during LU), send a BSSMAP Classmark Request to get it. To manage the intermission of the Classmark Request, add - msc_classmark_request_then_cipher_mode_cmd(), - state SUBSCR_CONN_S_WAIT_CLASSMARK_UPDATE, - event SUBSCR_CONN_E_CLASSMARK_UPDATE. From state AUTH_CIPH, switch to state WAIT_CLASSMARK_UPDATE. Once the BSSMAP Classmark Response, is received, switch back to SUBSCR_CONN_S_AUTH_CIPH and re-initiate Ciphering Mode. To be able to re-enter the Ciphering Mode algo decision, factor it out into msc_geran_set_cipher_mode(). Rationale: In the following commit, essentially we stopped supporting A5/3 ciphering: commit 71330720b6efdda2fcfd3e9c0cb45f89e32e5670 "MSC: Intersect configured A5 algorithms with MS-supported ones" Change-Id: Id124923ee52a357cb7d3e04d33f585214774f3a3 A5/3 was no longer supported because from that commit on, we strictly checked the MS-supported ciphers, but we did not have Classmark 2 available during Location Updating. This patch changes that: when Classmark 2 is missing, actively request it by a BSSMAP Classmark Request; continue Ciphering only after the Response. Always request missing Classmark, even if a lesser cipher were configured available. If the Classmark Update response fails to come in, cause an attach failure. Instead, we could attempt to use a lesser cipher that is also enabled. That is left as a future feature, should that become relevant. I think it's unlikely. Technically, we could now end up requesting a Classmark Updating both during LU (vlr_lu_fsm) and CM Service/Paging Response (proc_arq_fsm), but in practice the only time we lack a Classmark is: during Location Updating with A5/3 enabled. A5/1 support is indicated in CM1 which is always available, and A5/3 support is indicated in CM2, which is always available during CM Service Request as well as Paging Response. So this patch has practical relevance only for Location Updating. For networks that permit only A5/3, this patch fixes Location Updating. For networks that support A5/3 and A5/1, so far we always used A5/1 during LU, and after this patch we request CM2 and likely use A5/3 instead. In msc_vlr_test_gsm_ciph, verify that requesting Classmark 2 for A5/3 works during LU. Also verify that the lack of a Classmark Response results in attach failure. In msc_vlr_test_gsm_ciph, a hacky unit test fakes a situation where a CM2 is missing during proc_arq_fsm and proves that that code path works, even though the practical relevance is currently zero. It would only become interesting if ciphering algorithms A5/4 and higher became relevant, because support of those would be indicated in Classmark 3, which would always require a Classmark Request. Related: OS#3043 Depends: I4a2e1d3923e33912579c4180aa1ff8e8f5abb7e7 (libosmocore) Change-Id: I73c7cb6a86624695bd9c0f59abb72e2fdc655131
2018-09-13 01:23:07 +00:00
return a_iface_tx_cipher_mode(conn, &ei, retrieve_imeisv);
}
/* VLR asks us to start using ciphering.
* (Keep non-static to allow regression testing on this function.) */
int msc_vlr_set_ciph_mode(void *msc_conn_ref,
bool umts_aka,
bool retrieve_imeisv)
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
{
struct ran_conn *conn = msc_conn_ref;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
struct vlr_subscr *vsub;
struct vlr_auth_tuple *tuple;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
if (!conn || !conn->vsub) {
LOGP(DMM, LOGL_ERROR, "Cannot send Ciphering Mode Command to"
" NULL conn/subscriber");
return -EINVAL;
}
vsub = conn->vsub;
tuple = vsub->last_tuple;
if (!tuple) {
LOGP(DMM, LOGL_ERROR, "subscr %s: Cannot send Ciphering Mode"
" Command: no auth tuple available\n",
vlr_subscr_name(vsub));
return -EINVAL;
}
switch (conn->via_ran) {
case OSMO_RAT_GERAN_A:
return ran_conn_geran_set_cipher_mode(conn, umts_aka, retrieve_imeisv);
case OSMO_RAT_UTRAN_IU:
#ifdef BUILD_IU
DEBUGP(DMM, "-> SECURITY MODE CONTROL %s\n",
vlr_subscr_name(conn->vsub));
return ranap_iu_tx_sec_mode_cmd(conn->iu.ue_ctx, &tuple->vec, 0, 1);
#else
LOGP(DMM, LOGL_ERROR, "Cannot send Security Mode Control over OSMO_RAT_UTRAN_IU,"
" built without Iu support\n");
return -ENOTSUP;
#endif
default:
break;
}
LOGP(DMM, LOGL_ERROR,
"%s: cannot start ciphering, unknown RAN type %d\n",
vlr_subscr_name(conn->vsub), conn->via_ran);
return -ENOTSUP;
}
void ran_conn_rx_sec_mode_compl(struct ran_conn *conn)
{
struct vlr_ciph_result vlr_res = {};
if (!conn || !conn->vsub) {
LOGP(DMM, LOGL_ERROR,
"Rx Security Mode Complete for invalid conn\n");
return;
}
DEBUGP(DMM, "<- SECURITY MODE COMPLETE %s\n",
vlr_subscr_name(conn->vsub));
vlr_res.cause = VLR_CIPH_COMPL;
vlr_subscr_rx_ciph_res(conn->vsub, &vlr_res);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
}
/* VLR informs us that the subscriber data has somehow been modified */
static void msc_vlr_subscr_update(struct vlr_subscr *subscr)
{
LOGVSUBP(LOGL_NOTICE, subscr, "VLR: update for IMSI=%s (MSISDN=%s, used=%d)\n",
subscr->imsi, subscr->msisdn, subscr->use_count);
refactor log ctx for vlr_subscr and ran_conn ran_conn_get_conn_id(): instead of a talloc allocated string, return a static buffer in ran_conn_get_conn_id(). So far this function had no callers. Refactor ran_conn_update_id() API: during early L3-Complete, when no subscriber is associated yet, update the FSM Id by the MI type seen in the L3 Complete message: ran_conn_update_id_from_mi(). Later on set the vsub and re-update. Call vlr.ops->subscr_update when the TMSI is updated, so that log context includes the TMSI from then on. Enrich context for vlr_subscr_name and ran_conn fi name. Include all available information in vlr_subscr_name(); instead of either IMSI or MSISDN or TMSI, print all of them when present. Instead of a short log, rather have more valuable context. A context info would now look like: Process_Access_Request_VLR(IMSI-901700000014706:MSISDN-2023:TMSI-0x08BDE4EC:GERAN-A-3:PAGING_RESP) It does get quite long, but ensures easy correlation of any BSSAP / IuCS messages with log output, especially if multiple subscribers are busy at the same time. Print TMSI and TMSInew in uppercase hexadecimal, which is the typical representation in the telecom world. When showing the RAN conn id GERAN_A-00000017 becomes GERAN-A-23 - We usually write the conn_id in decimal. - Leading zeros are clutter and might suggest hexadecimal format. - 'GERAN-A' and 'UTRAN-Iu' are the strings defined by osmo_rat_type_name(). Depends: I7798c3ef983c2e333b2b9cbffef6f366f370bd81 (libosmocore) Depends: Ica25919758ef6cba8348da199b0ae7e0ba628798 (libosmocore) Change-Id: I66a68ce2eb8957a35855a3743d91a86299900834
2019-01-03 01:32:14 +00:00
ran_conn_update_id_for_vsub(subscr);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
}
static void update_classmark(const struct gsm_classmark *src, struct gsm_classmark *dst)
{
if (src->classmark1_set) {
dst->classmark1 = src->classmark1;
dst->classmark1_set = true;
}
if (src->classmark2_len) {
dst->classmark2_len = src->classmark2_len;
memcpy(dst->classmark2, src->classmark2, sizeof(dst->classmark2));
}
if (src->classmark3_len) {
dst->classmark3_len = src->classmark3_len;
memcpy(dst->classmark3, src->classmark3, sizeof(dst->classmark3));
}
}
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
/* VLR informs us that the subscriber has been associated with a conn */
static int msc_vlr_subscr_assoc(void *msc_conn_ref,
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
struct vlr_subscr *vsub)
{
struct ran_conn *conn = msc_conn_ref;
OSMO_ASSERT(vsub);
if (conn->vsub) {
if (conn->vsub == vsub)
LOG_RAN_CONN(conn, LOGL_NOTICE, "msc_vlr_subscr_assoc(): conn already associated with %s\n",
vlr_subscr_name(vsub));
else {
LOG_RAN_CONN(conn, LOGL_ERROR, "msc_vlr_subscr_assoc(): conn already associated with a subscriber,"
" cannot associate with %s\n", vlr_subscr_name(vsub));
return -EINVAL;
}
}
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
conn->vsub = vlr_subscr_get(vsub);
OSMO_ASSERT(conn->vsub);
conn->vsub->cs.attached_via_ran = conn->via_ran;
/* In case we have already received Classmark Information before the VLR Subscriber was
* associated with the conn: merge the new Classmark into vsub->classmark. Don't overwrite valid
* vsub->classmark with unset classmark, though. */
update_classmark(&conn->temporary_classmark, &conn->vsub->classmark);
refactor log ctx for vlr_subscr and ran_conn ran_conn_get_conn_id(): instead of a talloc allocated string, return a static buffer in ran_conn_get_conn_id(). So far this function had no callers. Refactor ran_conn_update_id() API: during early L3-Complete, when no subscriber is associated yet, update the FSM Id by the MI type seen in the L3 Complete message: ran_conn_update_id_from_mi(). Later on set the vsub and re-update. Call vlr.ops->subscr_update when the TMSI is updated, so that log context includes the TMSI from then on. Enrich context for vlr_subscr_name and ran_conn fi name. Include all available information in vlr_subscr_name(); instead of either IMSI or MSISDN or TMSI, print all of them when present. Instead of a short log, rather have more valuable context. A context info would now look like: Process_Access_Request_VLR(IMSI-901700000014706:MSISDN-2023:TMSI-0x08BDE4EC:GERAN-A-3:PAGING_RESP) It does get quite long, but ensures easy correlation of any BSSAP / IuCS messages with log output, especially if multiple subscribers are busy at the same time. Print TMSI and TMSInew in uppercase hexadecimal, which is the typical representation in the telecom world. When showing the RAN conn id GERAN_A-00000017 becomes GERAN-A-23 - We usually write the conn_id in decimal. - Leading zeros are clutter and might suggest hexadecimal format. - 'GERAN-A' and 'UTRAN-Iu' are the strings defined by osmo_rat_type_name(). Depends: I7798c3ef983c2e333b2b9cbffef6f366f370bd81 (libosmocore) Depends: Ica25919758ef6cba8348da199b0ae7e0ba628798 (libosmocore) Change-Id: I66a68ce2eb8957a35855a3743d91a86299900834
2019-01-03 01:32:14 +00:00
ran_conn_update_id(conn);
return 0;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
}
static int msc_vlr_route_gsup_msg(struct vlr_subscr *vsub,
struct osmo_gsup_message *gsup_msg)
{
switch (gsup_msg->message_type) {
/* GSM 09.11 code implementing SS/USSD */
case OSMO_GSUP_MSGT_PROC_SS_REQUEST:
case OSMO_GSUP_MSGT_PROC_SS_RESULT:
case OSMO_GSUP_MSGT_PROC_SS_ERROR:
DEBUGP(DMSC, "Routed to GSM 09.11 SS/USSD handler\n");
return gsm0911_gsup_handler(vsub, gsup_msg);
/* GSM 04.11 code implementing MO SMS */
case OSMO_GSUP_MSGT_MO_FORWARD_SM_ERROR:
case OSMO_GSUP_MSGT_MO_FORWARD_SM_RESULT:
case OSMO_GSUP_MSGT_READY_FOR_SM_ERROR:
case OSMO_GSUP_MSGT_READY_FOR_SM_RESULT:
DEBUGP(DMSC, "Routed to GSM 04.11 MO handler\n");
return gsm411_gsup_mo_handler(vsub, gsup_msg);
/* GSM 04.11 code implementing MT SMS */
case OSMO_GSUP_MSGT_MT_FORWARD_SM_REQUEST:
DEBUGP(DMSC, "Routed to GSM 04.11 MT handler\n");
return gsm411_gsup_mt_handler(vsub, gsup_msg);
default:
LOGP(DMM, LOGL_ERROR, "No handler found for %s, dropping message...\n",
osmo_gsup_message_type_name(gsup_msg->message_type));
return -GMM_CAUSE_MSGT_NOTEXIST_NOTIMPL;
}
}
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
/* operations that we need to implement for libvlr */
static const struct vlr_ops msc_vlr_ops = {
.tx_auth_req = msc_vlr_tx_auth_req,
.tx_auth_rej = msc_vlr_tx_auth_rej,
.tx_id_req = msc_vlr_tx_id_req,
.tx_lu_acc = msc_vlr_tx_lu_acc,
.tx_lu_rej = msc_vlr_tx_lu_rej,
.tx_cm_serv_acc = msc_vlr_tx_cm_serv_acc,
.tx_cm_serv_rej = msc_vlr_tx_cm_serv_rej,
.set_ciph_mode = msc_vlr_set_ciph_mode,
.tx_common_id = msc_vlr_tx_common_id,
.tx_mm_info = msc_vlr_tx_mm_info,
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
.subscr_update = msc_vlr_subscr_update,
.subscr_assoc = msc_vlr_subscr_assoc,
.forward_gsup_msg = msc_vlr_route_gsup_msg,
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
};
/* Allocate net->vlr so that the VTY may configure the VLR's data structures */
int msc_vlr_alloc(struct gsm_network *net)
{
net->vlr = vlr_alloc(net, &msc_vlr_ops);
if (!net->vlr)
return -ENOMEM;
net->vlr->user_ctx = net;
return 0;
}
/* Launch the VLR, i.e. its GSUP connection */
int msc_vlr_start(struct gsm_network *net)
{
struct ipaccess_unit *ipa_dev;
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
OSMO_ASSERT(net->vlr);
ipa_dev = talloc_zero(net->vlr, struct ipaccess_unit);
ipa_dev->unit_name = "MSC";
ipa_dev->serno = net->msc_ipa_name; /* NULL unless configured via VTY */
ipa_dev->swversion = PACKAGE_NAME "-" PACKAGE_VERSION;
return vlr_start(ipa_dev, net->vlr, net->gsup_server_addr_str, net->gsup_server_port);
Use libvlr in libmsc (large refactoring) Original libvlr code is by Harald Welte <laforge@gnumonks.org>, polished and tweaked by Neels Hofmeyr <nhofmeyr@sysmocom.de>. This is a long series of trial-and-error development collapsed in one patch. This may be split in smaller commits if reviewers prefer that. If we can keep it as one, we have saved ourselves the additional separation work. SMS: The SQL based lookup of SMS for attached subscribers no longer works since the SQL database no longer has the subscriber data. Replace with a round-robin on the SMS recipient MSISDNs paired with a VLR subscriber RAM lookup whether the subscriber is currently attached. If there are many SMS for not-attached subscribers in the SMS database, this will become inefficient: a DB hit returns a pending SMS, the RAM lookup will reveal that the subscriber is not attached, after which the DB is hit for the next SMS. It would become more efficient e.g. by having an MSISDN based hash list for the VLR subscribers and by marking non-attached SMS recipients in the SMS database so that they can be excluded with the SQL query already. There is a sanity limit to do at most 100 db hits per attempt to find a pending SMS. So if there are more than 100 stored SMS waiting for their recipients to actually attach to the MSC, it may take more than one SMS queue trigger to deliver SMS for subscribers that are actually attached. This is not very beautiful, but is merely intended to carry us over to a time when we have a proper separate SMSC entity. Introduce gsm_subscriber_connection ref-counting in libmsc. Remove/Disable VTY and CTRL commands to create subscribers, which is now a task of the OsmoHLR. Adjust the python tests accordingly. Remove VTY cmd subscriber-keep-in-ram. Use OSMO_GSUP_PORT = 4222 instead of 2222. See I4222e21686c823985be8ff1f16b1182be8ad6175. So far use the LAC from conn->bts, will be replaced by conn->lac in Id3705236350d5f69e447046b0a764bbabc3d493c. Related: OS#1592 OS#1974 Change-Id: I639544a6cdda77a3aafc4e3446a55393f60e4050
2016-06-19 16:06:02 +00:00
}
struct msgb *gsm48_create_mm_serv_rej(enum gsm48_reject_value value)
{
struct msgb *msg;
struct gsm48_hdr *gh;
msg = gsm48_msgb_alloc_name("GSM 04.08 SERV REJ");
if (!msg)
return NULL;
gh = (struct gsm48_hdr *) msgb_put(msg, sizeof(*gh) + 1);
gh->proto_discr = GSM48_PDISC_MM;
gh->msg_type = GSM48_MT_MM_CM_SERV_REJ;
gh->data[0] = value;
return msg;
}
struct msgb *gsm48_create_loc_upd_rej(uint8_t cause)
{
struct gsm48_hdr *gh;
struct msgb *msg;
msg = gsm48_msgb_alloc_name("GSM 04.08 LOC UPD REJ");
if (!msg)
return NULL;
gh = (struct gsm48_hdr *) msgb_put(msg, sizeof(*gh) + 1);
gh->proto_discr = GSM48_PDISC_MM;
gh->msg_type = GSM48_MT_MM_LOC_UPD_REJECT;
gh->data[0] = cause;
return msg;
}
int gsm48_extract_mi(uint8_t *classmark2_lv, int length, char *mi_string, uint8_t *mi_type)
{
/* Check the size for the classmark */
if (length < 1 + *classmark2_lv)
return -1;
uint8_t *mi_lv = classmark2_lv + *classmark2_lv + 1;
if (length < 2 + *classmark2_lv + mi_lv[0])
return -2;
*mi_type = mi_lv[1] & GSM_MI_TYPE_MASK;
return gsm48_mi_to_string(mi_string, GSM48_MI_SIZE, mi_lv+1, *mi_lv);
}
int gsm48_paging_extract_mi(struct gsm48_pag_resp *resp, int length,
char *mi_string, uint8_t *mi_type)
{
static const uint32_t classmark_offset =
offsetof(struct gsm48_pag_resp, classmark2);
uint8_t *classmark2_lv = (uint8_t *) &resp->classmark2;
return gsm48_extract_mi(classmark2_lv, length - classmark_offset,
mi_string, mi_type);
}