To help split openbsc.git to separate MSC and SGSN repositories, place the
common Iu interface related code here in libosmo-ranap. Also apply various
improvements while moving (from intermittent code review).
The code depends on libosmo-ranap tightly. One reason to want this separate
from libosmo-ranap could be that it uses libosmo-sigtran, accepting an sccp
instance. However, including in libosmo-ranap is the simplest way to go. The
osmo-iuh build depends on libosmo-sigtran anyway because of OsmoHNBGW, and all
current users of libosmo-ranap also naturally link libosmo-sigtran already.
Apply prefix ranap_iu_ and RANAP_IU_ to allow smooth transition from the
openbsc.git iu_ to the libranap ranap_iu_ implementations.
Prune unneeded #include statements.
Instead of sccp_addr, store an rnc pointer in the ue_conn_ctx. To facilitate,
also:
- Move iu_rnc struct to iu_client.h (as ranap_iu_rnc).
- Instead of sccp_addr, pass rnc to ue_conn_ctx_alloc().
- Pass a local struct new_ue_conn_ctx containing the sccp_addr and conn_id up
the RANAP handling stack in case of an InitialUE message.
- Separate the InitialUE message handling from cn_ranap_handle_co(), by moving
to new and separate cn_ranap_handle_co_initial(), so we can still pass a
looked-up ue_conn_ctx to all other cn_ranap_handle_co() code paths.
- Allocate the ue_conn_ctx only in ranap_handle_co_initial_ue(), not as early
as before.
Note that we are not actually ever using the rnc pointer now present in
ue_conn_ctx. It could be used for more concise paging, to first page only the
RNC where we last saw the subscriber. So far we page all matching LAC/RACs.
Tweak error logging: use __func__ instead of writing the function names as
string constants.
In iu_client_vty.c:
- Move the asn.1 debug commands from logging over to the iu node. They are not
specific to the logging target. They could qualify for an entirely separate
'asn1' root node, but for simplicity place under 'iu'.
- Add the 'asn1' commands to ranap_iu_vty_config_write(), so far missing.
- remove the legacy "net." from a VTY error message, it is not known which name
the parent node of 'iu' has.
Depends: libosmo-sccp I85b46269dbe7909e52873ace3f720f6292a4516c,
libosmo-sccp Ie1aedd7894acd69ddc887cd65a8a0df4b888838c
Change-Id: I6a3f7ad15be03fb94689b4af6ccfa828c25f45c0
When receiving unitdata from the CN, verify that it is indeed coming from the
remote address that matches our CS/PS domain settings.
This patch came from an earlier stage where the is_ps out-parameter was
actually used. While it currently isn't, it doesn't hurt to leave it there.
Change-Id: I7190b4c3a05e8bac0eeffa1eab18c9e47429cb17
With a headroom of 256, the SGSN crashes with:
msgb(0x6e3b90): Not enough headroom msgb_push (256 < 264)
I'm not perfectly sure what amount of headroom is strictly necessary. The only
reason to pick 512 is that it is twice the amount of 256.
Change-Id: I9a193846902a0477af0873f78283c4f2bedaf5dd
This might help in resolving whether we are conforming to SCCP specs regarding
our calling/called address choices.
Change-Id: Ie91953f7e9033419257d556d3a9f14e125f5c60d
In rua_to_scu(), log domain indicator and remote address.
In rua_rx_init_connect(), add domain indicator to log.
Change-Id: Ibe510bf0aca5eadbc0ce6031d3026e770de6559f
libosmo-sigtran now has a "proper" SCCP/M3UA stack, so we can make our hnb-gw
3GPP compliant by switching from the old SUA code to the new universal SCCP
user API with support for (currently) M3UA and SUA.
Main changes:
Use one cn_link to STP: We will connect to the core network using an (Osmo)STP
instance that routes to MSC and SGSN, so we want one SCCP link instead of two.
The only difference between IuCS and IuPS is a different remote osmo_sccp_addr.
This has various effects through the messaging code; the patch is a bit larger
than I would like, but it is hard to separate out truly independent smaller
changes.
CS or PS domain was previously flagged in the separate cn_link, as ctx pointer
for two separate sccp_sap_up()s. Now there's just one such ctx, so determine
is_ps from the RANAP Domain Indicator, or from the conn's hnbgw_context_map:
- Add is_ps to context_map_alloc_by_hnb().
- To find a matching context, the RUA ID alone is no longer sufficient, also
match is_ps (possible optimization todo: separate lists).
We would send separate CS or PS Reset messages based on the cn_link, instead
send both CS and PS Reset at the same time for the single cn_link. This could
be adjusted to detect presence of MSC or SGSN instead.
Pending: adjust the VTY config to reflect that there is only one remote
address. Place a TODO comment for that.
Smaller changes:
rua_to_scu(): populate called and calling addresses for N_CONNECT and
N_UNITDATA.
Remove DSUA.
Don't build dummy_cn, which is still implemented on SUA. Mark todo to maybe
re-include it based on M3UA later.
In hnbgw_cnlink, place sccp related items in a separate sub-struct.
Do not keep an llist of cn_links, just have the one. Remove iteration and list
management.
Change jenkins script to build libosmo-sccp master.
Patch-by: hwelte, nhofmeyr
Change-Id: I8ac15fa2fd25bedb26297177e416976a5389b573
After merge of the sysmocom/sctp branch to master, it was forgotten to change
this build script to move along to master.
Change-Id: I473d1361b99ff0ff8adba9326b0e2dfbc48a3aef
Return NULL in the error code path, otherwise an uninitialized pointer
is returned and later accessed when a UE tries to register using the
cnlink.
Change-Id: I4f3e2e0680de3216e2e569958bd64f70dc30c2a3
The script is expected to be run using python 2.x, but nowadays some
distros are already using python 3 as default, which will fail to run
this script.
This change fixes compilation in my Archlinux box.
Change-Id: I6eb95351538a64f2b23d638824972818591b1b66
'bind' was changed to local-ip in 39ee926062
The 'logging all everything' is known to not work, use 'all debug' instead.
Fix indenting of the 'log' items.
Change-Id: Ie7828073a4b4a5555f418246b20588e903ef4135
The guaranteedBitrate lowers from 12200 to 6700, which is mimicking the values
found in a trace from a production 3G environment. So far we have no reason to
choose these values other than knowing that other operators seem to do this.
This came up while trying to fix voice RAB for the nano3G, but this patch had
no effect on that. Now that it's here, we might as well keep it.
Change-Id: Ia7eecca43d62a6a020466e9b8dc8b566ca988f9f
For IPv4 addresses, only 56bit of X.213 NSAP are used. The sysmocell5000
accepts such a field that is 56bits long, but the ip.access nano3G does not
(and crashes instead). Both work when zero-padded to 160bits size.
So far we used to send the IPv4 address as "raw" 32bits to the nano3G to avoid
the crash. With this zero padding we no longer need such a workaround and can
use identical config for both cells.
Change-Id: I070bbfe887ab93d08322df30571050a381d082d5
This is mimicking the IEs found in a trace from a production 3G environment. We
have no reason to add this other than knowing that other operators seem to do
this.
This came up while trying to fix voice RAB for the nano3G, but this patch had
no effect on that. Now that it's here, we might as well keep it.
Change-Id: I14c22b0befb308bac2eded662fe13a58c8478743
CN_DomainIndicator_t is a long, so use %l format.
Error was introduced in recent 2b5021fd8d.
Fixes: CID#157119
Change-Id: I22f49d45bc87897beb24a6b1eee6c83da3e57b0f
Basically copy-paste the Iuh local-ip and local-port code to provide
parameterization of the IuCS and IuPS remote addresses.
Add IUCS and IUPS nodes, enhance go_parent_cb and config writing accordingly.
Change-Id: I2c28977011009df4e1fa472290bbbc359e406971
A second level of depth will be added to the hnbgw node soon, which will need
explicit go-parent logic.
Change-Id: I8d1c18a396c215e8425ae49872b5c73316087d7d
For completeness' sake, add VTY command to set the local Iuh port, to go with
the command that sets the local IP.
Change-Id: I4b5e9fe9fcfa489069a0728d47899ef4a61f7ce5
Use the g_hnb_gw->config.iuh_local_ip directly, drop hnbgw_get_iuh_local_ip().
Change-Id: Ie91aea82ae5d128ad735a0857ea814b440c3232c
Suggested-by: hwelte
Prepare for parameterization of IuCS and IuPS addresses:
Conform internal variable naming to local-ip, local-port, remote-ip,
remote-port (instead of bind-ip).
Rename HNBGW_IUH_LOCAL_IP_DEFAULT to HNGGW_LOCAL_IP_DEFAULT to be more general
and move it to the top.
Move a function doc comment to the .c file.
Change-Id: Ice85941c978498e3ddf41d151248507e7f56cb5d
Properly initialize msgb talloc context in hnbgw and all tests, using the new
msgb_talloc_ctx_init().
test-ranap.c: since msgb talloc ctx is now in test_common_init(), remove msgb
talloc init here.
Change-Id: I807c799aff1239184728551ed77fdafa73bd683f
Explicitly set AC_CONFIG_AUX_DIR.
To reproduce the error avoided by this patch:
rm install-sh # in case it was already generated.
touch ../install-sh # yes, outside this source tree
autoreconf -fi
This will produce an error like
...
configure.ac:16: error: required file '../ltmain.sh' not found
configure.ac:5: installing '../missing'
src/Makefile.am: installing '../depcomp'
autoreconf: automake failed with exit status: 1
See also automake (vim `which automake`) and look for 'sub locate_aux_dir'.
Change-Id: I5aad87b7d4c043f890f4ae2bc5d68cbad6d2c5c7
Assume that cat-testlogs.sh from osmo-ci is installed in $HOME/osmo-ci/scripts,
and call from jenkins.sh upon 'make check' failure.
Change-Id: I6ee37ae638aa0b95222e5c0ec78474fab1f8cfae
In the course of a rebase in libosmo-sccp, osmo_sua_link_get_user_priv() was
renamed to osmo_sccp_link_get_user_priv(). Apply this rename here to fix the
build.
Change-Id: I09603f5eb983f17c1e4a1232967fc43c7ceea12a
The standard osmo VTY terminology is 'remote-ip', 'remote-port', 'local-ip',
'local-port'. Conform to that. osmo-hnbgw is so far not rolled out widely, so
it makes sense to do this now.
Change-Id: Ifda2653bf58044552a5f1477cd7008dec3fb9100
Add the option to allow UE Register Requests with a TMSI identity.
Add VTY command to enable this option, 'hnbap-allow-tmsi'.
Add hnbgw_tx_ue_register_acc_tmsi().
HNBGW so far keeps track of UEs that have registered, with their IMSI. When a
UE registers with only a TMSI, we obviously can't store an IMSI. However, since
we're so far never *using* the list of UEs in osmo-hnbgw, we might as well just
accept the TMSI registration and carry on as usual. All that is needed for
proper operation is a valid UE context.
This is aimed at the ip.access nano3G femto cell, as it apparently feeds
whichever identification the UE sends through to HNBAP (TMSI+LAI, pTMSI+RAI),
instead of an IMSI as expected. So far this caused failures and the need to
make the UE clear its TMSI (wait several minutes or attempt to subscribe to a
different network), so that UE registration switched back to IMSI. When simply
accepting the TMSI in osmo-hngw, no problems are apparent in our current code
state.
For example, a Samsung Galaxy S4 seems to send a UE_Identity_PR_tMSILAI (CS
identity), and a GT-I9100 seems to send a UE_Identity_PR_pTMSIRAI (PS identity)
upon first registration to the network.
Recording the IMSI in hnbgw: we could use the subscriber list during paging, to
page a UE on only its last seen HNB. On the other hand, it doesn't hurt to
anyway always page to all HNBs connected to osmo-hnbgw. The paging procedure
does include a page-to-all-HNBs in case the first HNB paging fails. But we must
be aware that UEs that register by TMSI will simply not have an IMSI recorded
in the list of UE contexts, so a lookup based on IMSI may fail.
Patch-by: Harald Welte <laforge@gnumonks.org>, me
Change-Id: I87bc1aa3e85815ded7ac1dbdca48f1680b468589
To prepare for an upcoming commit that accepts TMSI identification upon UE
Register Requests:
Add tmsi arg to ue_context_alloc().
Add ue_context_by_tmsi().
This is aimed at the ip.access nano3G femto cell, as it apparently feeds
whichever identification the UE sends through to HNBAP (TMSI+LAI, pTMSI+RAI),
instead of an IMSI as expected.
See the upcoming commit that enables accepting TMSI identities for further
detail.
Change-Id: I138458443319cc4cbea5ee7906cf5dd72d582130
This is aimed at the ip.access nano3G femto cell, as it apparently feeds
whichever identification the UE sends through to HNBAP (TMSI+LAI, pTMSI+RAI),
instead of an IMSI as expected.
Sending a proper registration reject speeds up the response seen on the UE and
avoids needless waiting.
See the upcoming commit that enables accepting TMSI identities for further
detail.
Change-Id: I03b69613e6ddd8a08d9358ffc2f74954c231fd2c
After libosmocore 55dc2edc89c1a85187ef8aafc09f7d922383231f which outputs
'telnet at <ip> <port>' from telnet_init_dynif(), there's no need to log the
telnet VTY bind here anymore.
Change-Id: Icd9e670c1d30c156f7bd5d0d34892150aeba95e9
Remove an #if 0 to properly include the port information (verified to work).
Adjust test expectations.
Change-Id: I45fb134959dea9bcdfbfd9d8a061e67c3cc80fb7
This rule is bad because it re-invokes $(MAKE), but it seems to fix the
parallel build. It should probably be done differently.
Change-Id: I8107e08e4c211f20d671f520bf6bab9356f3c90e
This came up while fixing 'make distcheck'; this is certainly not the easiest
way but it makes sense to have the headers in include/, like we do in openbsc.
The easy alternative might be to add -I$(top_srcdir)/src to src/Makefile.am.
Remove -I$(top_srcdir)/src from src/tests/Makefile.am, no longer needed.
Change-Id: I5a82e029dcdc4df0a60a31271a4883393fe59234
Alexander Couzens
Neels Hofmeyr
Daniel Willmann
Pau Espin
Harald Welte