This fixes both a GCC and a Coverity warning:
GCC:
gsm_subscriber.c: In function ‘subscr_expire_callback’:
gsm_subscriber.c:389:2: warning: format ‘%i’ expects argument of type ‘int’, but argument 8 has type ‘long long unsigned int’ [-Wformat]
Coverity:
CID 1040712
This is fixing a GCC and Coverity warning.
GCC:
sgsn_vty.c: At top level:
sgsn_vty.c:308:1: warning: ‘show_ggsn_cmd’ defined but not used [-Wunused-variable]
Coverity: CID 1040727
This is fixing a GCC and Coverity warning:
GCC:
sgsn_vty.c: In function ‘vty_dump_pdp’:
sgsn_vty.c:64:5: warning: ‘i’ may be used uninitialized in this function [-Wmaybe-uninitialized]
sgsn_vty.c:49:15: note: ‘i’ was declared here
Coverity: CID 1040706
The code was written like checking the return value of the
osmo_fd_register but the rc variable was not assigned for the
subsequent calls.
Fixes: Coverity CID 1040741
Original code was inconsistent about lengths and could lead to out
of bounds write. Lengths were also inconsistent with the TS 24.008.
Fixes: Coverity CID 1040714.
The support has been implemented for an old model, we were told that
newer versions would be made incompatible with OpenBSC. Ther are
various warnings in the code and coverity has found some new ones.
Just remove the code as we don't know of anyone using this code.
Coverity is not happy about it but it doesn't appear to be a real
issue as the data will not be NULL on the first call.
Addresses: Coverity CID 1040704
In case the specified trx number is not configured, do not crash
but return NULL from the function. The libosmo-abis library should
close the connection for us then.
In some situations (like MS reboot without prior DETACH or SGSN reboot
without prior MS detach), the LLC sequence numbers for UI mode could
be different on both sides.
The LLC spec unfortunately doesn't permit us to send something like a
FRMR in this case, but instructs us to silently discard the frame. At
that time the remote LLC entity will re-transmit the frame with the same
seqeunce number over and over again, which we will drop again and again.
The mthod used now will keep track of the last received UI sequence
number. If that number is retransmitted for three times in a row, then
we accept this sequence number and recover from that point on.
system_id is a char array, which makes the NULL check a tautology,
so we should check with strlen if the string is empty or not.
This is fixing the "write" command of VTY that would otherwise
create a config file that can not be parsed.
If an ESME has the dcs_transparent config flag, then the TP-DCS
of MO-SMS is transparently passed to SMPP and not converted to SMPP
specific data_coding values.
This is needed in cases where ESMEs actually care about the exact
TP-DCS, as the conversion from TP-DCS to SMPP data_coding is not
bijective.
There are multiple ways how the TS 03.38 TP-DCS can indicate 8bit or
7bit messages. SMPP has it's own data coding specification, which is
different from TS 03.38.
However, some SMPP ESMEs want to be able to have fine-grained control
over the TP-DCS indicated in the TPDU header. If such values like 0xF6
are used in SMPP, we now transparently pass them on to the GSM side.
The RQNT message might not contain a 'S:' line with the actual tone
to play. Instead of calling the callback with the 0 as tone just leave
early.
Example:
X: 6B9519B88F0
R: D/[0-9#*](N), G/ft, fxr/t38
A channel that fails to send an ACK/NACK/REL within the four second
timeout is now marked as broken. In case the release comes in late
it will be ignored. The lchan should already been freed and for now
we don't want to trust the channel.
In the future we might want to send a "release" for a channel that
got activated ack late and just set the state to none on a channel
that is released acked late.
The late ACK and REL has been tested with two manual tests using the
fakebts. The channels remained blocked even after having received
the ACK message here. The NACK case has not been manually tested.
Assume that there is a bsc_gsmnet and modify the ipaccess-config to
provide this symbol. If a bsc_gsmmnet is not available when linked
the linker will complain and fail.
E.g. give an error like this:
../../src/libbsc/libbsc.a(bts_ipaccess_nanobts.o): In function `ipaccess_sign_link_up':
src/libbsc/bts_ipaccess_nanobts.c:550: undefined reference to `bsc_gsmnet'
Make sure that the bts_ipa_nm_sig_cb is only installed once. In
a nanoBTS + sysmoBTS network the _start method will be called
for each mode leading to the handler being registered twice.
Make sure that there is only one handler registered by unregistering
a previously installed handler. The osmo_signal_unregister_handler
will remove 0 or 1 handlers per invocation and we only add one
handler in this file so it is always balanced.
The nanoBTS code is trying to find a struct gsm_bts based on
the ipaccess_gsmnet and the ipaccess_unit data. The pointer is
not initialized in the case of a sysmoBTS leading to a classic
NULL pointer dereference.
Move the feature init into the _init method. This way we can
re-use the start code of the nanoBTS. This ensures that the
ipaccess_gsmnet pointer is properly initialized and that the
signal handlers are installed.
Extend the status message and send LAC/CI as part of the status
message. It is using TV to allow sending more fields in the feature.
We only need to encode the data and this is why there is no tlv
description yet.
The bsc_nat.h is included by common_vty.c so we may not used
sccp_types.h in the bsc_nat.h header file. Move the callstats
to a new file and include it where it is needed.
Find the Cell Identifier from the Complete Layer3 Information and
store it for future reference. We could begin to verify that the
LAC/CI used really belongs to the BSC.
The name sccp_connection is used in the osmo-sccp code, sccp_connections
was used in the NAT for tracking a sccp_connection. Rename it so it is
obvious that the struct belongs to the nat.
The rename was done with sed:
$ sed -i s,"struct sccp_connections","struct nat_sccp_connection",g \
include/openbsc/*.h src/osmo-bsc_nat/* tests/*/*
Prepend the international number with a '+' and then do the normal
re-writing on it. There are a couple of ways to handle this:
\+([0-9]), \+[0-9][0-9]([0-9]), \+49([0-9])
Add a test case for the international re-write based on an already
internationalized number.
Limit the amount of pending DLCX responses to three times the amount
of available endpoints. Currently all MGCP messages are sent and handled
in sequence.
We want to send a TRAP with the MGCP statistics from the NAT and
the connected BSC. The BSC endpoint can be either released because
of a DLCX from the MGCP CallAgent or the SCCP Connection release on
the A-link.
This is why we need to queue the statistics when the deleting the
endpoint on the BSC. The processing is continued once the response
arrives. This code assumes that the response of the DLCX will be sent
by the remote side. The current amount of outstanding responses can be
seen on the VTY. This assumption is based on the fact that the BSC has
already responded to the CRCX and maybe to the MDCX.
The MGCP RFC is bended to prefix the transaction identifier with "nat-"
to easily detect the response and hand it to the handler. This will
then parse the response and generate the TRAP. The current version is
v1. We assume that the transaction space is big enough and we will
not re-assign the transaction identifier too early.
The sysmobts is now having a SAPI queue with all pending SAPI operations
on the BTS. Add the llist_head to the lchan and make sure it is initialized
by the shared code.
The token was compared with the configured one but only up to a
user supplied length. Compare the token sizes and then use memcmp
for the actual comparison to make sure to compare the right ammount
of characters.
There is no unit-test but there should be one.
It was possible that the tmp was strduped but not freed, e.g. when
the number of commas was not right. It would evenutally be freed at
the time the cmd is freed thanks to the talloc hierachy.
An ESME can now be configured in the VTY to enable osmocom-extensions,
which will add vendor-specific SMPP TLVs for RxLev/RxQual/ARFCN/IMEI and
transmit power to the SMPP DELIVER-SM message type.
As bsc_gsmnet is NULL at the time we call smpp_openbsc_init(),
we later run into segfaults with subscribers that don't have a
subscr->net set.
However, we cannot delay smpp_openbsc_init() until after
bsc_bootstrap_network(), as we then fail to parse the SMPP specific
VTY/config file options...
This patch adds a new VTY command "ip.access rsl-ip A.B.C.D" at the
BTS level. If you set this IP address, the BTS will be instructed to
establish the RSL link to the indiciated IP address, rather than using
the same as for the OML link (default).
Use "ip.access rsl-ip 0" to disable the feature.
Inside the SI1 rest_octets we will need to indicate if the ARFCN
is band 1800 or 1900. If the BTS is either 850 or 1900 we assume
we are running a PCS network, otherwise it is a DCS network.
The band indicator is not documented in GSM 04.08 but it is in the
GSM 05.14 version 6.1.0 Release 1997.
abis_rsl.c:332:23: warning: 'memset' call operates on objects of type 'struct rsl_ie_chan_mode'
while the size is based on a different type 'struct rsl_ie_chan_mode *'
[-Wsizeof-pointer-memaccess]
memset(cm, 0, sizeof(cm));
~~ ^~
abis_rsl.c:332:23: note: did you mean to dereference the argument to 'sizeof' (and multiply it
by the number of elements)?
memset(cm, 0, sizeof(cm));
^~
1 warning generated.
The paging structure is still initialized lazily and we attempt
to traverse it before it is ready. The crash was discovered by
Katerina. Removing the lazy initialization will take a bit of work
as the gsm_data_shared is used by the osmo-bts and the OpenBSC but
paging is different.
Katerina pointed out that some nodes are not fully documented and
proposed some messages. The token/timeout messages were correct, I
have modified the other messages. I removed the full-stop from the
PING/PONG documentation as we are normally not using a full sentence.
when the SMS code was moved into libosmocore, its logging prefix was
changed from 'sms' to 'lsms', which breaks existing config files.
This introduces a deprecated/hidden vty command to make sure those
config files are still parsed OK, and will simply print a warning
message about the config needing some update.
The message was corrupt at several points. They are fixed now and
successfully tested.
A default T3122 timer value of 10 is defined by default now. If set to 0,
the reject message will not be sent. Note that when using existing configs
with T3122 value set to 0.
The RF lock excluded BTS was not paged at all. Now forward the
paging message to the handler and call a function that will check
if this LAC can be paged right now. Introduce a new paging method
that allows to page on a dedicated bts, refactor the code to use
this method for paging.
When introducing the exclude for the BTS lock the RF stayed up but
all connections were immediately released. Optionally pass the BTS
as second parameter and check the exclude bit.
Tested-with: rf-lock-exclude/RFLockExcludeTest.st
If after release has been sent, the call control layer waits for:
- reception of release complete
- or timeout
- or release of transaction (due to radio link failure)
In this case, an MNCC_REL_CNF is sent to upper layer. The callref must
still exist, so the upper layer can handle this confirm.
If transaction is destroyed, but callref still exists, the
mncc_release_ind function is called. If the upper layer already sent an
MNCC_REL_REQ, the state N19 was entered. In this case it expects an
MNCC_REL_CNF.
For PCS1900 the SI1 does not contain the ARFCN of the serving cell.
This is because the arfcn2band method will return GSM_BAND_1800 and
not GSM_BAND_1900. The academic fix would be to set the ARFCN_PCS bit
but this would require increasing the bitvector sizes from 1024/8 to
(0x8000 + 1024) / 8. This would increase the storage size for each
bitvector by three.
It is not possible to have DCS1800 and PCS1900 in one network so we
can avoid increasing the memory usage and check if the ARFCN resolved
to 1800 and then check if the BTS is a 1900 BTS and then claim that
this is compatible.
The code predates the TLV parser and we were parsing the RLM from the
wrong offset. In general we were using the length of the TLV which
happened to be equal to the T200 indication.
After consulting the RLM cuases not every of them should generate a
BSC_RLLR_IND_ERR_IND as these are forwarded to the MSC as a SAPI reject
right now.
TLV parsing now generates this due a bug in the osmo-bts code:
abis_rsl.c:1605 (bts=0,trx=0,ts=2,ss=0) SAPI=0 <0000> abis_rsl.c:1547 (bts=0,trx=0,ts=2,ss=0) ERROR INDICATION cause=Fraeme not implemented
For short IP failures we want the RF to stay up and wait for
the re-connect but in case the A-link is gone too long it is
good to switch off the RF and wait for commands to enable it
again.
The issue can be reproduced by typing the following 9 or more times.
OpenBSC> subscriber id 2 sms sender id 2 send bla
For some unknown reason the phone sends us a CP-ERROR for a transaction
identifier we have allocated and used but don't remember. Due the way
we use the SMC/SMR we 'establish' the machine and this results in a CP-ACK
being sent out. But the CP-ERROR is not having the content we want for
an establish so we send out a RP-ERROR. This will result in a CP-ERROR
because the phone does not know the transaction...
Avoid the issue by checking the direction of the transaction. If we do
not know the transaction and it is supposed to be allocated by us then
just ignore it and do not create a new transaction.
Handle the mr_config request and set the AMR multirate config for
the given MSC. Initialize the mr_config with the AMR5.9 default we
have been using until now.
When we are asked to route calls on a local link and
the link is not available we would crash when trying
to send a packet over a deadline. When we have decided
to move a connection it is guranteed that the current
SCCP connection will vanish, we either migrate to another
MSC or the RSL/subscriber connection will be closed.
Inspect the CC Setup messages and if the dialed number is matching
the regexp of the local MSC the connection will be rerouted. The
original MSC will get a GSM0808 CLEAR REQUEST, a new connection with
a CC Setup message will be opened.
For USSD we remember that it is a supplementary service but this
means we sent no CM Service Reject down to the subscriber. Treat
NAT_CON_TYPE_CM_SERV_REQ and NAT_CON_TYPE_SSA the same and send
a cm service reject.
Do the auth check in bsc_nat_filter_sccp_cr, remove the cause from
the signature again. For the bsc_nat_filter_dt restructure the flow
but leave the auth inside the id response message.
Return 1 when the IMSI has been extracted as indicator for running
the auth check. 1 has not been used before and is safe to be used
as this indicator.
In preparation for another kind of black-list allow the filter code
to decide how the connection should be rejected. Introduce a new struct
that will carry the reject causes for certain operations.
Set the subscriber expiry timeout to twice the duration of the location
update period and provide functions subscr_expire() and
db_subscriber_expire() to mark subscribers offline that have missed two
location update periods.
This patch increases the DB revision to 3, so the hlr will be
incompatible with prior versions.
We should allow 0 for T3212 as well to disable the location update
period. In that case we will need a way to indicate that in the
database.
libcommon: Default to 30min location update period
libbsc: Limit VTY value for periodic update and disallow the value 0
According to GSM 04.08 Table 10.5.33 "The value 0 is used for infinite
timeout value i.e. periodic updating shall not be used within the cell."
This was the default value until now, but the code that deals with
expiring inactive subscribers in the next commit can't handle that case
so this remains a TODO for now.
The actual command implementation was already for this, but some
bad vty parsing prevented optional range arguments from working
properly.
Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
This was reported by Kevin when he was testing handover. The problem
is the order of the signal handlers for S_ABISIP_CRCX_ACK. Right now
the handover signal handler is called before the one inside the libmsc
gsm_04_08.c. This means S_HANDOVER_ACK is signalled _before_ there is a
rtp socket created for the channel. The result is that the MDCX will
never be sent and the called will not be properly switched _after_ the
handover detection.
I do not want to play with the order of signal handlers, remove the
CRCX ack handling from the handover_logic.c and force the NITB (and
later the BSC) to check if the lchan is involved with a handover and
do the switching in there. This means right now we do what two signal
handlers did in one.
Reproduced and tested with the FakeBTS Handover test.
Log message:
<0004> abis_rsl.c:1954 (bts=1,trx=0,ts=3,ss=0) IPAC_CRCX_ACK ...
<000c> gsm_04_08.c:1400 no RTP socket for new_lchan
<001a> rtp_proxy.c:533 rtp_socket_create(): success
<001a> rtp_proxy.c:615 rtp_socket_bind(rs=0x48703c8, IP=0.0.0.0): ...
In case of handover (but probably on RACH) we would send a RLL for
SAPI=0 even if this SAPI was never established. After we have released
all SAPI>0 locally check that SAPI=0 is established and if not release
the rf channel directly.
T3109 is started when the SACCH is deactivated. It is stopped when
the phones sends the DISC/UA/UM on LAPDm for the main signalling
link. In case of timeout the abnormal release procedure will be
initiated. Make sure to not issue the SACCH Deactivate twice to
avoid confusing the equipment.
This is still not fully spec compliant. In case of a timeout the
abnormal release handling will be started which involves starting
T3111+2. The error handling should be split out of the rf channel
release method, e.g. lchan_release should be called and check if
the channel release was already initiated.
If the CHAN ACTIV is NACKED we set the state backto NONE. This is
problematic as our channel allocator will allocate from the front
or from the back and if the channel is early in the list it might
cause permanent failures. Introduce a BROKEN state and use it when
the channel activation is failing for an unknown reason. Copy the
cause so it can be inspected later.
Deactivate the SACCH and release all SAPIs locally as of GSM 04.08. Add
documentation to the code and explain what will happen as part of the
release process.
* Release all channels with SAPI > 0 with the "local end release"
(as of NOTE 1 of GSM 04.08).
* No need to wait for all SAPIs to be torn down and the normal
REL_IND/REL_CONF will call rsl_handle_release and the channel
should be released.
* Update the documentation
Put the idiom that sets the lchan state to none and respecting the
error state into a shared method. This way the special handling for
the abnormal case is just in one place.
Avoid printing error messages when paging while the BTS is down. In
case a BTS is going down it is best to just let the timers expire
normally. We can not expire them right away as multiple BTS could be
paged and we do not want to interfere with that. There is no need to
stop the queue right now.
If subscriber A is calling B and has sent a CC Setup message we will
allocate the MO and MT transaction and link them together. When the
BTS or the lchan is failing the BSC API will send a clear request,
as part of the clear request all pending transactions will be released.
As part of taking down the transaction, the remote leg will be informed
and will send a MNCC_REL_REQ. This results in a call to trans_free. The
llist_for_each_entry_safe does not handle removing other elements from
the list and we would segfault.
One way to fix this is to move the transaction list into the subscriber
connection. This might require to create the subscriber connection for
MT handling earlier. Otherwise one could have one transaction list inside
the subscriber connection and a global list for MT- transactions.
This has been reported and analyzed by Tobias Engel. The IMSI Detach
is dispatched as part of the complete layer3 message. I had patched
the code to release the anchor and call msc_release_connection to
release the connection as fast as possible (otherwise the anchor would
trigger in a couple of seconds).
With commit 70ae5d3000 I made this more
generic to release the connection immediately if there are no operations,
no transaction and no silent call. This leads to the subscriber connection
being released twice and eventually causing a segfault. Remove the
msc_release_connection invocation from the IMSI Detach code as the connection
will be taken down by the BSC API.
This has been tested using the FakeBTS and an IMSI Detach message. The
channel is released immediately and the nitb does not crash.
GCC 3.x on PowerPC correctly highlights that the code is fishy.
Re-reading the RFC 3550 shows that we should subtract it and then
we are in the 16bit range. The probation and re-sync code is still
missing.
GCC:
mgcp/mgcp_network.c:200: warning: comparison is always true due to limited range of data type
MGCP is used over UDP and a response might be lost. The MGCP RFC
asks for keeping a list of responses and then using the previous
response to answer a duplicate request. I tried to conserve memory
and just wanted to remember the last transaction identifier and
result-code and re-generate the result from that. This made the
code look bad and this is why the entire response will now be stored.
It sadly increases the memory usage but can not be avoided at this
time.
Remove the msg->l3h pointer for the RQNT callback as strtok has
modified the content of it.
Rebased, tested and fixed by Holger Freyther. Release the transaction
only once the SMC is asking for the release and set the cb's to NULL
to catch a use after free early.
This has been rebased and fixed by Holger Freyther. The change of
the debug area was split out in a previous commit and the is_mt was
put back into the transaction code.
The transaction is now freed from the RELEASE_REQ sent by the SMC
layer and not inside the error path. When clearing the SMC instance
we also clear the callbacks.
This should and does happen as part of the trans_free/msc_release_connection
code. There is no easy way to determine that the lchan is now 'free' for other
things.. Let the transaction code sort this out. This code just needs to make
sure that transactions are always freed.
This is more a work around and one still needs to implement a
proper dispatch on the opening of the connection. If there is no
operation left, no transaction and no silent call, close down the
channel.
Use the code that is shipped inside the libosmogsm library. Right now
the signature (besides the static) and the implementation is the same.
This makes using the libosmogsm SMC code more easy in the near future.
For the gsm340_gen_oa we are now using a small wrapper to generate the
proper type and numbering plan.
For some reason, libsmpp34 is too smart to zero out the entire structure
to which it is unpacking. This introduces an ugly wrapper macro to
work around. This needs discussion with the libsmpp34 maintainer.
Move to the control command handling out of the main file into
a dedicated module. There are still some calls embedded into the
main code but it will be moved soon.
Use a usec timestamp for the local time. The seconds to usec will
swap over to the lower bits but this appears to be correct. The
CLOCK_MONOTONIC is used to fulfill the RFC 3550 requirement even
if it is a bit slower than the gettimeofday.
Make sure to initialize transit in a way that the first transit
time will be 0. Otherwise the jitter will contain the difference
of the localtime and the remote time.
Calculate the expected packages and packet loss as of RFC 3550.
The values should be clamped but our packet loss counter is 32
bits and not 24 and we should clamp at other values but I am
waiting for some issues first before dealing with that.
This is missing the probation and the dealing with a remote
restart. For the remote restart we will simply write a log
statement as this is unlikely to happen during a call or if
it does happen the call will be taken down by the BSC anyway.
Align the naming inside the mgcp_rtp_state with the naming inside
the 'source' struct of the appendix. Make first_seq_no/base_seq
a uint16_t. This is removing rules for alignments and reduces the
struct from 40 bytes to 36.
Count the received octets. This is encouraged by the MGCP specification.
Use a 32bit counter that is good enough for more than 12 hours of a EFR
call. This limit is good enough for the current configuration.
The RFC 3435 specifies a different formula for calculating the lost
packages. It involves the number of received packages and the delta
of the sequence number.
The previous code didn't work as expected. The trx and dst pointer
are located in an union and in the case of the Abis code the dst
is used to point to the signalling link timeslot and not the TRX.
The is_ipaccess_bts always returned false because the dst was casted
to a trx while it was no trx.
This fix was tested with the nack_test/NACKTest.st of the test repo.
The test cases were failing on 64bit systems because the sizeof
code operated on the pointer size which is 8 and longer than the
size that was intended to be used for comparing it.
It is a bad idea to detach a subscriber. The subscriber will not
be reachable until the next periodic updating cycle. In case we have
too many failed deliveries we will need to reduce the period for the
LU and implement a subscriber purging task.
This is a preparation for the 29C3 and a problem Jolly experience with
his type writer system.
I saw the old copy of the "Appendix J" code too late and I have
discovered some quirks and I am more familar with my implementation.
Most noticable 'w' only needs to be as big as the input arfcn but
requires the 'w' to be initialized. The power_of_2 implementation
differs as well (mine matches the output of wirehsark).
The f0 could be chosen in a better way but right now picking
the lower bound is the easiest. It is not clear if to use
modulo if the range is chosen in the middle. This can be improved
in the future. Right now I have no bit fiddling for range128, 256
and 1024 as I was running out of time.
alpha=0 (the new value) doesn't reduce MS transmission power during GPRS
as much as we did with alpha=10. This is to optimize for coverage and
to keep GPRS working at all cost, and not care about MS battery life
time or uplink interference in surrounding cells.
FIXME: This should be made configurable via the VTY and the normal
default (unless configured otherwise by vty/config file) should be '6'.
In order to keep mobile at PACCH as long as possible the timer T3192 is
set to 1500ms. This reduces the probablity of long lasting assignment
process on CCCH for subsequent downlink TBFs.
Holger Hans Peter Freyther
Harald Welte
Alexander Huemer
Andreas Eversberg
Sylvain Munaut