osmo-hlr/src/hlr_ussd.c

721 lines
22 KiB
C
Raw Normal View History

/* OsmoHLR SS/USSD implementation */
/* (C) 2018 Harald Welte <laforge@gnumonks.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 <osmocom/core/talloc.h>
#include <osmocom/core/timer.h>
#include <osmocom/gsm/gsup.h>
#include <osmocom/gsm/gsm0480.h>
#include <osmocom/gsm/protocol/gsm_04_80.h>
#include <stdint.h>
#include <string.h>
#include <errno.h>
#include <osmocom/hlr/hlr.h>
#include <osmocom/hlr/hlr_ussd.h>
#include <osmocom/hlr/gsup_server.h>
#include <osmocom/hlr/gsup_router.h>
#include <osmocom/hlr/logging.h>
#include <osmocom/hlr/db.h>
/***********************************************************************
* core data structures expressing config from VTY
***********************************************************************/
struct hlr_euse *euse_find(struct hlr *hlr, const char *name)
{
struct hlr_euse *euse;
llist_for_each_entry(euse, &hlr->euse_list, list) {
if (!strcmp(euse->name, name))
return euse;
}
return NULL;
}
struct hlr_euse *euse_alloc(struct hlr *hlr, const char *name)
{
struct hlr_euse *euse = euse_find(hlr, name);
if (euse)
return NULL;
euse = talloc_zero(hlr, struct hlr_euse);
euse->name = talloc_strdup(euse, name);
euse->hlr = hlr;
llist_add_tail(&euse->list, &hlr->euse_list);
return euse;
}
void euse_del(struct hlr_euse *euse)
{
llist_del(&euse->list);
talloc_free(euse);
}
struct hlr_ussd_route *ussd_route_find_prefix(struct hlr *hlr, const char *prefix)
{
struct hlr_ussd_route *rt;
llist_for_each_entry(rt, &hlr->ussd_routes, list) {
if (!strcmp(rt->prefix, prefix))
return rt;
}
return NULL;
}
struct hlr_ussd_route *ussd_route_prefix_alloc_int(struct hlr *hlr, const char *prefix,
const struct hlr_iuse *iuse)
{
struct hlr_ussd_route *rt;
if (ussd_route_find_prefix(hlr, prefix))
return NULL;
rt = talloc_zero(hlr, struct hlr_ussd_route);
rt->prefix = talloc_strdup(rt, prefix);
rt->u.iuse = iuse;
llist_add_tail(&rt->list, &hlr->ussd_routes);
return rt;
}
struct hlr_ussd_route *ussd_route_prefix_alloc_ext(struct hlr *hlr, const char *prefix,
struct hlr_euse *euse)
{
struct hlr_ussd_route *rt;
if (ussd_route_find_prefix(hlr, prefix))
return NULL;
rt = talloc_zero(hlr, struct hlr_ussd_route);
rt->prefix = talloc_strdup(rt, prefix);
rt->is_external = true;
rt->u.euse = euse;
llist_add_tail(&rt->list, &hlr->ussd_routes);
return rt;
}
void ussd_route_del(struct hlr_ussd_route *rt)
{
llist_del(&rt->list);
talloc_free(rt);
}
static struct hlr_ussd_route *ussd_route_lookup_for_req(struct hlr *hlr, const struct ss_request *req)
{
const uint8_t cgroup = req->ussd_data_dcs >> 4;
const uint8_t lang = req->ussd_data_dcs & 0x0f;
char ussd_code[GSM0480_USSD_7BIT_STRING_LEN];
struct hlr_ussd_route *rt;
ussd_code[0] = '\0';
/* We support only the Coding Group 0 (GSM 7-bit default alphabeet). In fact,
* the USSD request is usually limited to [*#0-9], so we don't really need to
* support other coding groups and languages. */
switch (cgroup) {
case 0:
/* The Language is usually set to '1111'B (unspecified), but some UEs
* are known to indicate '0000'B (German). */
if (lang != 0x0f) {
LOGP(DSS, LOGL_NOTICE, "USSD DataCodingScheme (0x%02x): "
"the Language is usually set to 15 (unspecified), "
"but the request indicates %u - ignoring this\n",
req->ussd_data_dcs, lang);
/* do not abort, attempt to decode as if it was '1111'B */
}
gsm_7bit_decode_n_ussd(&ussd_code[0], sizeof(ussd_code),
req->ussd_data, (req->ussd_data_len * 8) / 7);
break;
default:
LOGP(DSS, LOGL_ERROR, "USSD DataCodingScheme (0x%02x): "
"Coding Group %u is not supported, expecting Coding Group 0\n",
req->ussd_data_dcs, cgroup);
return NULL;
}
llist_for_each_entry(rt, &hlr->ussd_routes, list) {
if (!strncmp(ussd_code, rt->prefix, strlen(rt->prefix))) {
LOGP(DSS, LOGL_DEBUG, "Found %s '%s' (prefix '%s') for USSD "
"Code '%s'\n", rt->is_external ? "EUSE" : "IUSE",
rt->is_external ? rt->u.euse->name : rt->u.iuse->name,
rt->prefix, ussd_code);
return rt;
}
}
LOGP(DSS, LOGL_DEBUG, "Could not find Route for USSD Code '%s'\n", ussd_code);
return NULL;
}
/***********************************************************************
* handling functions for individual GSUP messages
***********************************************************************/
#define LOGPSS(ss, lvl, fmt, args...) \
LOGP(DSS, lvl, "%s/0x%08x: " fmt, (ss)->imsi, (ss)->session_id, ## args)
struct ss_session {
/* link us to hlr->ss_sessions */
struct llist_head list;
/* imsi of this session */
char imsi[OSMO_IMSI_BUF_SIZE];
/* ID of this session (unique per IMSI) */
uint32_t session_id;
/* state of the session */
enum osmo_gsup_session_state state;
/* time-out when we will delete the session */
struct osmo_timer_list timeout;
/* is this USSD for an external handler (EUSE): true */
bool is_external;
union {
/* external USSD Entity responsible for this session */
struct hlr_euse *euse;
/* internal USSD Entity responsible for this session */
const struct hlr_iuse *iuse;
} u;
/* subscriber's vlr_number
* MO USSD: originating MSC's vlr_number
* MT USSD: looked up once per session and cached here */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
struct osmo_ipa_name vlr_name;
/* we don't keep a pointer to the osmo_gsup_{route,conn} towards the MSC/VLR here,
* as this might change during inter-VLR hand-over, and we simply look-up the serving MSC/VLR
* every time we receive an USSD component from the EUSE */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
struct osmo_gsup_req *initial_req_from_ms;
struct osmo_gsup_req *initial_req_from_euse;
};
struct ss_session *ss_session_find(struct hlr *hlr, const char *imsi, uint32_t session_id)
{
struct ss_session *ss;
llist_for_each_entry(ss, &hlr->ss_sessions, list) {
if (!strcmp(ss->imsi, imsi) && ss->session_id == session_id)
return ss;
}
return NULL;
}
void ss_session_free(struct ss_session *ss)
{
osmo_timer_del(&ss->timeout);
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
if (ss->initial_req_from_ms)
osmo_gsup_req_free(ss->initial_req_from_ms);
if (ss->initial_req_from_euse)
osmo_gsup_req_free(ss->initial_req_from_euse);
llist_del(&ss->list);
talloc_free(ss);
}
static void ss_session_timeout(void *data)
{
struct ss_session *ss = data;
LOGPSS(ss, LOGL_NOTICE, "SS Session Timeout, destroying\n");
/* FIXME: should we send a ReturnError component to the MS? */
ss_session_free(ss);
}
struct ss_session *ss_session_alloc(struct hlr *hlr, const char *imsi, uint32_t session_id)
{
struct ss_session *ss;
OSMO_ASSERT(!ss_session_find(hlr, imsi, session_id));
ss = talloc_zero(hlr, struct ss_session);
OSMO_ASSERT(ss);
OSMO_STRLCPY_ARRAY(ss->imsi, imsi);
ss->session_id = session_id;
/* Schedule self-destruction timer */
osmo_timer_setup(&ss->timeout, ss_session_timeout, ss);
if (g_hlr->ncss_guard_timeout > 0)
osmo_timer_schedule(&ss->timeout, g_hlr->ncss_guard_timeout, 0);
llist_add_tail(&ss->list, &hlr->ss_sessions);
return ss;
}
/***********************************************************************
* handling functions for encoding SS messages + wrapping them in GSUP
***********************************************************************/
/* Resolve the target MSC by ss->imsi and send GSUP message. */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
static int ss_gsup_send_to_ms(struct ss_session *ss, struct osmo_gsup_server *gs, struct osmo_gsup_message *gsup)
{
struct hlr_subscriber subscr = {};
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
struct msgb *msg;
int rc;
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
if (ss->initial_req_from_ms) {
/* Use non-final osmo_gsup_req_respond() to not deallocate the ss->initial_req_from_ms */
osmo_gsup_req_respond(ss->initial_req_from_ms, gsup, false, false);
return 0;
}
msg = osmo_gsup_msgb_alloc("GSUP USSD FW");
rc = osmo_gsup_encode(msg, gsup);
if (rc) {
LOGPSS(ss, LOGL_ERROR, "Failed to encode GSUP message\n");
msgb_free(msg);
return rc;
}
/* Use vlr_number as looked up by the caller, or look up now. */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
if (!ss->vlr_name.len) {
rc = db_subscr_get_by_imsi(g_hlr->dbc, ss->imsi, &subscr);
if (rc < 0) {
LOGPSS(ss, LOGL_ERROR, "Cannot find subscriber, cannot route GSUP message\n");
msgb_free(msg);
return -EINVAL;
}
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
osmo_ipa_name_set_str(&ss->vlr_name, subscr.vlr_number);
}
/* Check for empty string (all vlr_number strings end in "\0", because otherwise gsup_route_find() fails) */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
if (ss->vlr_name.len <= 1) {
LOGPSS(ss, LOGL_ERROR, "Cannot send GSUP message, no VLR number stored for subscriber\n");
msgb_free(msg);
return -EINVAL;
}
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
LOGPSS(ss, LOGL_DEBUG, "Tx SS/USSD to VLR %s\n", osmo_ipa_name_to_str(&ss->vlr_name));
return osmo_gsup_send_to_ipa_name(gs, &ss->vlr_name, msg);
}
static int ss_tx_to_ms(struct ss_session *ss, enum osmo_gsup_message_type gsup_msg_type,
struct msgb *ss_msg)
{
struct osmo_gsup_message resp;
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
int rc;
resp = (struct osmo_gsup_message) {
.message_type = gsup_msg_type,
.session_id = ss->session_id,
.session_state = ss->state,
};
OSMO_STRLCPY_ARRAY(resp.imsi, ss->imsi);
if (ss_msg) {
resp.ss_info = msgb_data(ss_msg);
resp.ss_info_len = msgb_length(ss_msg);
}
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
rc = ss_gsup_send_to_ms(ss, g_hlr->gs, &resp);
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
msgb_free(ss_msg);
return rc;
}
#if 0
static int ss_tx_reject(struct ss_session *ss, int invoke_id, uint8_t problem_tag,
uint8_t problem_code)
{
struct msgb *msg = gsm0480_gen_reject(invoke_id, problem_tag, problem_code);
LOGPSS(ss, LOGL_NOTICE, "Tx Reject(%u, 0x%02x, 0x%02x)\n", invoke_id,
problem_tag, problem_code);
OSMO_ASSERT(msg);
ss->state = OSMO_GSUP_SESSION_STATE_END;
return ss_tx_to_ms(ss, OSMO_GSUP_MSGT_PROC_SS_RESULT, msg);
}
#endif
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
static int ss_tx_to_ms_error(struct ss_session *ss, uint8_t invoke_id, uint8_t error_code)
{
struct msgb *msg = gsm0480_gen_return_error(invoke_id, error_code);
LOGPSS(ss, LOGL_NOTICE, "Tx ReturnError(%u, 0x%02x)\n", invoke_id, error_code);
OSMO_ASSERT(msg);
ss->state = OSMO_GSUP_SESSION_STATE_END;
return ss_tx_to_ms(ss, OSMO_GSUP_MSGT_PROC_SS_RESULT, msg);
}
static int ss_tx_to_ms_ussd_7bit(struct ss_session *ss, uint8_t invoke_id, const char *text)
{
struct msgb *msg = gsm0480_gen_ussd_resp_7bit(invoke_id, text);
LOGPSS(ss, LOGL_INFO, "Tx USSD '%s'\n", text);
OSMO_ASSERT(msg);
return ss_tx_to_ms(ss, OSMO_GSUP_MSGT_PROC_SS_RESULT, msg);
}
/***********************************************************************
* Internal USSD Handlers
***********************************************************************/
#include <osmocom/hlr/db.h>
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
static int handle_ussd_own_msisdn(struct ss_session *ss,
const struct osmo_gsup_message *gsup, const struct ss_request *req)
{
struct hlr_subscriber subscr;
char buf[GSM0480_USSD_7BIT_STRING_LEN+1];
int rc;
ss->state = OSMO_GSUP_SESSION_STATE_END;
rc = db_subscr_get_by_imsi(g_hlr->dbc, ss->imsi, &subscr);
switch (rc) {
case 0:
if (strlen(subscr.msisdn) == 0)
snprintf(buf, sizeof(buf), "You have no MSISDN!");
else
snprintf(buf, sizeof(buf), "Your extension is %s", subscr.msisdn);
ss_tx_to_ms_ussd_7bit(ss, req->invoke_id, buf);
break;
case -ENOENT:
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
ss_tx_to_ms_error(ss, req->invoke_id, GSM0480_ERR_CODE_UNKNOWN_SUBSCRIBER);
break;
case -EIO:
default:
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
ss_tx_to_ms_error(ss, req->invoke_id, GSM0480_ERR_CODE_SYSTEM_FAILURE);
break;
}
return 0;
}
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
static int handle_ussd_own_imsi(struct ss_session *ss,
const struct osmo_gsup_message *gsup, const struct ss_request *req)
{
char buf[GSM0480_USSD_7BIT_STRING_LEN+1];
snprintf(buf, sizeof(buf), "Your IMSI is %s", ss->imsi);
ss->state = OSMO_GSUP_SESSION_STATE_END;
ss_tx_to_ms_ussd_7bit(ss, req->invoke_id, buf);
return 0;
}
/* This handler just keeps the session idle unless the guard timer expires. */
static int handle_ussd_test_idle(struct ss_session *ss,
const struct osmo_gsup_message *gsup,
const struct ss_request *req)
{
char buf[GSM0480_USSD_7BIT_STRING_LEN + 1];
snprintf(buf, sizeof(buf), "Keeping your session idle, it will expire "
"at most in %u seconds.", g_hlr->ncss_guard_timeout);
ss->state = OSMO_GSUP_SESSION_STATE_CONTINUE;
ss_tx_to_ms_ussd_7bit(ss, req->invoke_id, buf);
return 0;
}
static const struct hlr_iuse hlr_iuses[] = {
{
.name = "own-msisdn",
.handle_ussd = handle_ussd_own_msisdn,
},
{
.name = "own-imsi",
.handle_ussd = handle_ussd_own_imsi,
},
{
.name = "test-idle",
.handle_ussd = handle_ussd_test_idle,
},
};
const struct hlr_iuse *iuse_find(const char *name)
{
unsigned int i;
for (i = 0; i < ARRAY_SIZE(hlr_iuses); i++) {
const struct hlr_iuse *iuse = &hlr_iuses[i];
if (!strcmp(name, iuse->name))
return iuse;
}
return NULL;
}
/***********************************************************************
* handling functions for individual GSUP messages
***********************************************************************/
static bool ss_op_is_ussd(uint8_t opcode)
{
switch (opcode) {
case GSM0480_OP_CODE_PROCESS_USS_DATA:
case GSM0480_OP_CODE_PROCESS_USS_REQ:
case GSM0480_OP_CODE_USS_REQUEST:
case GSM0480_OP_CODE_USS_NOTIFY:
return true;
default:
return false;
}
}
/* is this GSUP connection an EUSE (true) or not (false)? */
static bool peer_name_is_euse(const struct osmo_cni_peer_id *peer_name)
{
if (peer_name->type != OSMO_CNI_PEER_ID_IPA_NAME)
return false;
if (peer_name->ipa_name.len <= 5)
return false;
return strncmp((char *)(peer_name->ipa_name.val), "EUSE-", 5) == 0;
}
static struct hlr_euse *euse_by_name(const struct osmo_cni_peer_id *peer_name)
{
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
if (!peer_name_is_euse(peer_name))
return NULL;
/* above peer_name_is_euse() ensures this: */
OSMO_ASSERT(peer_name->type == OSMO_CNI_PEER_ID_IPA_NAME);
return euse_find(g_hlr, (const char*)(peer_name->ipa_name.val)+5);
}
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
static int handle_ss(struct ss_session *ss, bool is_euse_originated, const struct osmo_gsup_message *gsup,
const struct ss_request *req)
{
uint8_t comp_type = gsup->ss_info[0];
LOGPSS(ss, LOGL_INFO, "SS CompType=%s, OpCode=%s\n",
gsm0480_comp_type_name(comp_type), gsm0480_op_code_name(req->opcode));
/**
* FIXME: As we don't store any SS related information
* (e.g. call forwarding preferences) in the database,
* we don't handle "structured" SS requests at all.
*/
LOGPSS(ss, LOGL_NOTICE, "Structured SS requests are not supported, rejecting...\n");
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
ss_tx_to_ms_error(ss, req->invoke_id, GSM0480_ERR_CODE_FACILITY_NOT_SUPPORTED);
return -ENOTSUP;
}
/* Handle a USSD GSUP message for a given SS Session received from VLR or EUSE */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
static int handle_ussd(struct ss_session *ss, bool is_euse_originated, const struct osmo_gsup_message *gsup,
const struct ss_request *req)
{
uint8_t comp_type = gsup->ss_info[0];
struct msgb *msg_out;
LOGPSS(ss, LOGL_INFO, "USSD CompType=%s, OpCode=%s '%s'\n",
gsm0480_comp_type_name(comp_type), gsm0480_op_code_name(req->opcode),
req->ussd_text);
if ((ss->is_external && !ss->u.euse) || !ss->u.iuse) {
LOGPSS(ss, LOGL_NOTICE, "USSD for unknown code '%s'\n", req->ussd_text);
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
ss_tx_to_ms_error(ss, req->invoke_id, GSM0480_ERR_CODE_SS_NOT_AVAILABLE);
return 0;
}
if (is_euse_originated) {
/* Received from EUSE, Forward to VLR */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
/* Need a non-const osmo_gsup_message, because sending might modify some (routing related?) parts. */
struct osmo_gsup_message forward = *gsup;
ss_gsup_send_to_ms(ss, g_hlr->gs, &forward);
} else {
/* Received from VLR (MS) */
if (ss->is_external) {
/* Forward to EUSE */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
struct osmo_ipa_name euse_name;
struct osmo_gsup_conn *conn;
osmo_ipa_name_set_str(&euse_name, "EUSE-%s", ss->u.euse->name);
conn = gsup_route_find_by_ipa_name(g_hlr->gs, &euse_name);
if (!conn) {
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
LOGPSS(ss, LOGL_ERROR, "Cannot find conn for EUSE %s\n",
osmo_ipa_name_to_str(&euse_name));
ss_tx_to_ms_error(ss, req->invoke_id, GSM0480_ERR_CODE_SYSTEM_FAILURE);
} else {
msg_out = osmo_gsup_msgb_alloc("GSUP USSD FW");
osmo_gsup_encode(msg_out, gsup);
osmo_gsup_conn_send(conn, msg_out);
}
} else {
/* Handle internally */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
ss->u.iuse->handle_ussd(ss, gsup, req);
/* Release session if the handler has changed its state to END */
if (ss->state == OSMO_GSUP_SESSION_STATE_END)
ss_session_free(ss);
}
}
return 0;
}
/* this function is called for any SS_REQ/SS_RESP messages from both the MSC/VLR side as well
* as from the EUSE side */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
void rx_proc_ss_req(struct osmo_gsup_req *gsup_req)
{
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
struct hlr *hlr = g_hlr;
struct ss_session *ss;
struct ss_request req = {0};
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
const struct osmo_gsup_message *gsup = &gsup_req->gsup;
/* Remember whether this function should free the incoming gsup_req: if it is placed as ss->initial_req_from_*,
* do not free it here. If not, free it here. */
struct osmo_gsup_req *free_gsup_req = gsup_req;
bool is_euse_originated = peer_name_is_euse(&gsup_req->source_name);
LOGP(DSS, LOGL_DEBUG, "%s/0x%08x: Process SS (%s)\n", gsup->imsi, gsup->session_id,
osmo_gsup_session_state_name(gsup->session_state));
if (gsup_req->source_name.type != OSMO_CNI_PEER_ID_IPA_NAME) {
LOGP(DSS, LOGL_ERROR, "%s/0x%082x: Unable to process SS request: Unsupported GSUP peer id type%s\n",
gsup->imsi, gsup->session_id,
osmo_cni_peer_id_type_name(gsup_req->source_name.type));
osmo_gsup_req_respond_err(gsup_req, GMM_CAUSE_PROTO_ERR_UNSPEC, "error processing SS request");
return;
}
/* decode and find out what kind of SS message it is */
if (gsup->ss_info && gsup->ss_info_len) {
if (gsm0480_parse_facility_ie(gsup->ss_info, gsup->ss_info_len, &req)) {
LOGP(DSS, LOGL_ERROR, "%s/0x%082x: Unable to parse SS request: %s\n",
gsup->imsi, gsup->session_id,
osmo_hexdump(gsup->ss_info, gsup->ss_info_len));
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
osmo_gsup_req_respond_err(gsup_req, GMM_CAUSE_INV_MAND_INFO, "error parsing SS request");
return;
}
} else if (gsup->session_state != OSMO_GSUP_SESSION_STATE_END) {
LOGP(DSS, LOGL_ERROR, "%s/0x%082x: Missing SS payload for '%s'\n",
gsup->imsi, gsup->session_id,
osmo_gsup_session_state_name(gsup->session_state));
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
osmo_gsup_req_respond_err(gsup_req, GMM_CAUSE_INV_MAND_INFO, "missing SS payload");
return;
}
switch (gsup->session_state) {
case OSMO_GSUP_SESSION_STATE_BEGIN:
/* Check for overlapping Session ID usage */
if (ss_session_find(hlr, gsup->imsi, gsup->session_id)) {
LOGP(DSS, LOGL_ERROR, "%s/0x%08x: BEGIN with non-unique session ID!\n",
gsup->imsi, gsup->session_id);
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
osmo_gsup_req_respond_err(gsup_req, GMM_CAUSE_INV_MAND_INFO, "BEGIN with non-unique session ID");
return;
}
ss = ss_session_alloc(hlr, gsup->imsi, gsup->session_id);
if (!ss) {
LOGP(DSS, LOGL_ERROR, "%s/0x%08x: Unable to allocate SS session\n",
gsup->imsi, gsup->session_id);
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
osmo_gsup_req_respond_err(gsup_req, GMM_CAUSE_NET_FAIL, "Unable to allocate SS session");
return;
}
/* Get IPA name from VLR conn and save as ss->vlr_number */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
if (!is_euse_originated) {
ss->initial_req_from_ms = gsup_req;
free_gsup_req = NULL;
OSMO_ASSERT(gsup_req->source_name.type == OSMO_CNI_PEER_ID_IPA_NAME); /* checked above */
ss->vlr_name = gsup_req->source_name.ipa_name;
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
} else {
ss->initial_req_from_euse = gsup_req;
free_gsup_req = NULL;
}
if (ss_op_is_ussd(req.opcode)) {
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
if (is_euse_originated) {
/* EUSE->VLR: MT USSD. EUSE is known ('conn'), VLR is to be resolved */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
ss->u.euse = euse_by_name(&gsup_req->source_name);
} else {
/* VLR->EUSE: MO USSD. VLR is known ('conn'), EUSE is to be resolved */
struct hlr_ussd_route *rt;
rt = ussd_route_lookup_for_req(hlr, &req);
if (rt) {
if (rt->is_external) {
ss->is_external = true;
ss->u.euse = rt->u.euse;
} else if (rt) {
ss->is_external = false;
ss->u.iuse = rt->u.iuse;
}
} else {
if (hlr->euse_default) {
ss->is_external = true;
ss->u.euse = hlr->euse_default;
}
}
}
/* dispatch unstructured SS to routing */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
handle_ussd(ss, is_euse_originated, &gsup_req->gsup, &req);
} else {
/* dispatch non-call SS to internal code */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
handle_ss(ss, is_euse_originated, &gsup_req->gsup, &req);
}
break;
case OSMO_GSUP_SESSION_STATE_CONTINUE:
ss = ss_session_find(hlr, gsup->imsi, gsup->session_id);
if (!ss) {
LOGP(DSS, LOGL_ERROR, "%s/0x%08x: CONTINUE for unknown SS session\n",
gsup->imsi, gsup->session_id);
osmo_gsup_req_respond_err(gsup_req, GMM_CAUSE_MSGT_INCOMP_P_STATE,
"CONTINUE for unknown SS session");
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
return;
}
/* Reschedule self-destruction timer */
if (g_hlr->ncss_guard_timeout > 0)
osmo_timer_schedule(&ss->timeout, g_hlr->ncss_guard_timeout, 0);
if (ss_op_is_ussd(req.opcode)) {
/* dispatch unstructured SS to routing */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
handle_ussd(ss, is_euse_originated, &gsup_req->gsup, &req);
} else {
/* dispatch non-call SS to internal code */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
handle_ss(ss, is_euse_originated, &gsup_req->gsup, &req);
}
break;
case OSMO_GSUP_SESSION_STATE_END:
ss = ss_session_find(hlr, gsup->imsi, gsup->session_id);
if (!ss) {
LOGP(DSS, LOGL_ERROR, "%s/0x%08x: END for unknown SS session\n",
gsup->imsi, gsup->session_id);
osmo_gsup_req_respond_err(gsup_req, GMM_CAUSE_MSGT_INCOMP_P_STATE,
"END for unknown SS session");
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
return;
}
/* SS payload is optional for END */
if (gsup->ss_info && gsup->ss_info_len) {
if (ss_op_is_ussd(req.opcode)) {
/* dispatch unstructured SS to routing */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
handle_ussd(ss, is_euse_originated, &gsup_req->gsup, &req);
} else {
/* dispatch non-call SS to internal code */
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
handle_ss(ss, is_euse_originated, &gsup_req->gsup, &req);
}
}
ss_session_free(ss);
break;
default:
LOGP(DSS, LOGL_ERROR, "%s/0x%08x: Unknown SS State %d\n", gsup->imsi,
gsup->session_id, gsup->session_state);
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
break;
}
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
if (free_gsup_req)
osmo_gsup_req_free(free_gsup_req);
}
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
void rx_proc_ss_error(struct osmo_gsup_req *req)
{
1/2: refactor: add and use lu_fsm, osmo_gsup_req, osmo_ipa_name These are seemingly orthogonal changes in one patch, because they are in fact sufficiently intertwined that we are not willing to spend the time to separate them. They are also refactoring changes, unlikely to make sense on their own. ** lu_fsm: Attempting to make luop.c keep state about incoming GSUP requests made me find shortcomings in several places: - since it predates osmo_fsm, it is a state machine that does not strictly enforce the order of state transitions or the right sequence of incoming events. - several places OSMO_ASSERT() on data received from the network. - modifies the subscriber state before a LU is accepted. - dead code about canceling a subscriber in a previous VLR. That would be a good thing to actually do, which should also be trivial now that we record vlr_name and sgsn_name, but I decided to remove the dead code for now. To both step up the LU game *and* make it easier for me to integrate osmo_gsup_req handling, I decided to create a lu_fsm, drawing from my, by now, ample experience of writing osmo_fsms. ** osmo_gsup_req: Prepare for D-GSM, where osmo-hlr will do proxy routing for remote HLRs / communicate with remote MSCs via a proxy: a) It is important that a response that osmo-hlr generates and that is sent back to a requesting MSC contains all IEs that are needed to route it back to the requester. Particularly source_name must become destination_name in the response to be able to even reach the requesting MSC. Other fields are also necessary to match, which were so far taken care of in individual numerous code paths. b) For some operations, the response to a GSUP request is generated asynchronously (like Update Location Request -> Response, or taking the response from an EUSE, or the upcoming proxying to a remote HLR). To be able to feed a request message's information back into the response, we must thus keep the request data around. Since struct osmo_gsup_message references a lot of external data, usually with pointers directly into the received msgb, it is not so trivial to pass GSUP message data around asynchronously, on its own. osmo_gsup_req is the combined solution for both a and b: it keeps all data for a GSUP message by taking ownership of the incoming msgb, and it provides an explicit API "forcing" callers to respond with osmo_gsup_req_respond(), so that all code paths trivially are definitely responding with the correct IEs set to match the request's routing (by using osmo_gsup_make_response() recently added to libosmocore). Adjust all osmo-hlr code paths to use *only* osmo_gsup_req to respond to incoming requests received on the GSUP server (above LU code being one of them). In fact, the same should be done on the client side. Hence osmo_gsup_req is implemented in a server/client agnostic way, and is placed in libosmo-gsupclient. As soon as we see routing errors in complex GSUP setups, using osmo_gsup_req in the related GSUP client is likely to resolve those problems without much thinking required beyond making all code paths use it. libosmo-gsupclient is hence added to osmo-hlr binary's own library dependencies. It would have been added by the D-GSM proxy routing anyway, we are just doing it a little sooner. ** cni_peer_id.c / osmo_ipa_name: We so far handle an IPA unit name as pointer + size, or as just pointer with implicit talloc size. To ease working with GSUP peer identification data, I require: - a non-allocated storage of an IPA Name. It brings the drawback of being size limited, but our current implementation is anyway only able to handle MSC and SGSN names of 31 characters (see struct hlr_subscriber). - a single-argument handle for IPA Name, - easy to use utility functions like osmo_ipa_name_to_str(), osmo_ipa_name_cmp(), and copying by simple assignment, a = b. Hence this patch adds a osmo_ipa_name in cni_peer_id.h and cni_peer_id.c. Heavily used in LU and osmo_gsup_req. Depends: libosmocore Id9692880079ea0f219f52d81b1923a76fc640566 Change-Id: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
2019-11-20 01:36:45 +00:00
LOGP(DSS, LOGL_NOTICE, "%s/0x%08x: Process SS ERROR (%s)\n", req->gsup.imsi, req->gsup.session_id,
osmo_gsup_session_state_name(req->gsup.session_state));
osmo_gsup_req_free(req);
}