Hi Holger,
After trying to checkout and build the OpenBSC using command git checkout -b on-waves/bsc-master origin/bsc-master, we got error as
fatal: git checkout: updating paths is incompatible with switching branches. Did you intend to checkout 'origin/bsc-master' which can not be resolved as commit?
So, may be we are not correct on the path or something else. can you please advice, as we are not able to find bsc_msc_ip binary file.
On Thu, Sep 30, 2010 at 9:50 AM, Holger Hans Peter Freyther < holger@freyther.de> wrote:
On 09/30/2010 03:04 PM, tejas oza wrote:
Hi Guys,
We are implementing the separate MSC module and we will be using the
openBSC,
and we are mainly concerned for SMS services. So, I will be highly
obliged if
you can provide us the documentation on openBSC code so that we can
understand
the messages expected on separate MSC side from open BSC and the messages
to
be sent from MSC to openBSC side.
In the on-waves/bsc-master branch you will find a bsc_msc_ip binary that will be able to connect to a real MSC through somthing that is coined SCCP-Lite (SCCP encapsulated inside the ipaccess protocol).
If you look at bsc_hack it has no connection to anything but the BTS, so the BSC/MSC/VLR/HLR functionality is implemented in one binary.
On 09/30/2010 05:40 PM, tejas oza wrote:
Hi Holger,
After trying to checkout and build the OpenBSC using command git checkout -b on-waves/bsc-master origin/bsc-master, we got error as
fatal: git checkout: updating paths is incompatible with switching branches. Did you intend to checkout 'origin/bsc-master' which can not be resolved as commit?
So, may be we are not correct on the path or something else. can you please advice, as we are not able to find bsc_msc_ip binary file.
well... it is origin/on-waves/bsc-master and not origin/bsc-master....
After running the command git checkout -b origin/on-waves/bsc-master , we are getting fatal: You are on a branch yet to be born
On Thu, Sep 30, 2010 at 11:56 AM, Holger Hans Peter Freyther < holger@freyther.de> wrote:
On 09/30/2010 05:40 PM, tejas oza wrote:
Hi Holger,
After trying to checkout and build the OpenBSC using command git checkout -b on-waves/bsc-master origin/bsc-master, we got error as
fatal: git checkout: updating paths is incompatible with switching
branches.
Did you intend to checkout 'origin/bsc-master' which can not be resolved
as
commit?
So, may be we are not correct on the path or something else. can you
please
advice, as we are not able to find bsc_msc_ip binary file.
well... it is origin/on-waves/bsc-master and not origin/bsc-master....
Hello.
On Thu, 2010-09-30 at 13:13, tejas oza wrote:
After running the command git checkout -b origin/on-waves/bsc-master , we are getting fatal: You are on a branch yet to be born
I would suggest to search for a GIT tutorial and read it. This problem has nothing to do with OpenBSC development. Hint: After -b follows the name for the local branch in which you want to checkout the one specified.
Also it would be nice if you could stop top posting on this mailling list.
regards Stefan Schmidt
Hi Stefan,
I managed to fix the problem.Thanks for your support. And now onwards I will post related to OpenBSC stuffs on this mailing list. Sorry for any churns.
On Thu, Sep 30, 2010 at 1:39 PM, Stefan Schmidt stefan@datenfreihafen.orgwrote:
Hello.
On Thu, 2010-09-30 at 13:13, tejas oza wrote:
After running the command git checkout -b origin/on-waves/bsc-master , we are getting fatal: You are on a branch yet to be born
I would suggest to search for a GIT tutorial and read it. This problem has nothing to do with OpenBSC development. Hint: After -b follows the name for the local branch in which you want to checkout the one specified.
Also it would be nice if you could stop top posting on this mailling list.
regards Stefan Schmidt
Hi Holger,
We are going to build MSC using 'Erlang' which will communicate with OpenBSC 'C' code. Since we will be dealing with SMS services only, so during MO operation our MSC should get messages from OpenBSC and for MT operation we should pass the messages to OpenBSC. As we will follow SS7 protocol for communication of MSC to HLR and MSC to SMSC and we will be dealing with SCCP.
We are concerned about Open BSC and MSC communication for SMS service and we need to know from OpenBSC code that what particular messages are used for SMS transfer.
So can you please advice us to get started.
On Thu, Sep 30, 2010 at 10:58 PM, Holger Hans Peter Freyther < holger@freyther.de> wrote:
On 09/30/2010 08:09 PM, tejas oza wrote:
Hi Stefan,
I managed to fix the problem.Thanks for your support. And now onwards I
will
post related to OpenBSC stuffs on this mailing list. Sorry for any
churns.
Which MSC are you using? Does it support SCCP-lite out of the box?
On 10/01/2010 07:26 PM, tejas oza wrote:
Hi Holger,
We are going to build MSC using 'Erlang' which will communicatewith OpenBSC 'C' code. Since we will be dealing with SMS services only, so during MO operation our MSC should get messages from OpenBSC and for MT operation we should pass the messages to OpenBSC. As we will follow SS7 protocol for communication of MSC to HLR and MSC to SMSC and we will be dealing with SCCP.
Will you be able to release the resulting code a GPL or GPL compatible?
We are concerned about Open BSC and MSC communication for SMS service and we need to know from OpenBSC code that what particular messages are used for SMS transfer.
There is nothing SMS specific (besides using SAPI=3). Consult GSM08.08 for the messages between the MSC and the BSC.
Hi Holger,
If we will make any changes to OpenBSC code then we will release it as a GPL. And regarding our MSC code, we got the permission from company that we can release it as a GPL or GPL Compatible under the Licence for inhouse and testing use.
We gone through SCCP and now we are going to start implementing SCCP in Erlang and we gone through the SCCP files in Open BSC code. We are confused that if we have to implement all SCCP function same as in OpenBSC in MSC to communicate.
Please advice.
On Fri, Oct 1, 2010 at 1:38 PM, Holger Hans Peter Freyther < holger@freyther.de> wrote:
On 10/01/2010 07:26 PM, tejas oza wrote:
Hi Holger,
We are going to build MSC using 'Erlang' which willcommunicate
with OpenBSC 'C' code. Since we will be dealing with SMS services only,
so
during MO operation our MSC should get messages from OpenBSC and for MT operation we should pass the messages to OpenBSC. As we will follow SS7 protocol for communication of MSC to HLR and MSC to SMSC and we will be dealing with SCCP.
Will you be able to release the resulting code a GPL or GPL compatible?
We are concerned about Open BSC and MSC communication for SMS service and
we
need to know from OpenBSC code that what particular messages are used for
SMS
transfer.
There is nothing SMS specific (besides using SAPI=3). Consult GSM08.08 for the messages between the MSC and the BSC.
On 10/04/2010 10:48 PM, tejas oza wrote:
Hi Holger,
If we will make any changes to OpenBSC code then we will release it as a GPL. And regarding our MSC code, we got the permission from company that we can release it as a GPL or GPL Compatible under the Licence for inhouse and testing use.
It would be nice if you could publish code as soon as possible as I don't want to give free support for non free software.
We gone through SCCP and now we are going to start implementing SCCP in Erlang and we gone through the SCCP files in Open BSC code. We are confused that if we have to implement all SCCP function same as in OpenBSC in MSC to communicate.
The easiest is to write a C Node and wrap the libsccp.a
Hi Holger,
We will publish our code as soon as we produced it.
As you said we made C node and wrapped libsccp.a. After running bsc_msc_ip we are getting following error.
*ozmank@ozmank-desktop:~/libosmocore/openbsc/openbsc/src$ ./bsc_msc_ip 130.238.15.219 58620 Bootstraping the network. Sending GSM08.08 reset. mi_e1_line_update could not open socket Address family not supported by protocol full talloc report on 'sccp' (total 29 bytes in 2 blocks) struct sccp_data_callback contains 28 bytes in 1 blocks (ref 0) 0x9ca12a8*
Can you please advice.
On Mon, Oct 4, 2010 at 5:21 PM, Holger Hans Peter Freyther < holger@freyther.de> wrote:
On 10/04/2010 10:48 PM, tejas oza wrote:
Hi Holger,
If we will make any changes to OpenBSC code then we will release it as a
GPL.
And regarding our MSC code, we got the permission from company that we
can
release it as a GPL or GPL Compatible under the Licence for inhouse and testing use.
It would be nice if you could publish code as soon as possible as I don't want to give free support for non free software.
We gone through SCCP and now we are going to start implementing SCCP in
Erlang
and we gone through the SCCP files in Open BSC code. We are confused that
if
we have to implement all SCCP function same as in OpenBSC in MSC to
communicate.
The easiest is to write a C Node and wrap the libsccp.a
On 10/07/2010 05:12 PM, tejas oza wrote:
Hi Holger,
We will publish our code as soon as we produced it.
As you said we made C node and wrapped libsccp.a. After running bsc_msc_ip we are getting following error.
*ozmank@ozmank-desktop:~/libosmocore/openbsc/openbsc/src$ ./bsc_msc_ip 130.238.15.219 58620 Bootstraping the network. Sending GSM08.08 reset. mi_e1_line_update could not open socket Address family not supported by protocol full talloc report on 'sccp' (total 29 bytes in 2 blocks) struct sccp_data_callback contains 28 bytes in 1 blocks (ref 0) 0x9ca12a8*
Can you please advice.
Well,
what do you think that this error could be? Which kind of BTS do you attempt to use? OpenBSC supports the Siemens BS11 and the ipaccess nanoBTS, with a configuration file (that is documented in the wiki) you define which BTS to use... you have obviously created a openbsc.cfg otherwise it would have complained about a missing one.
Hi Holger,
Right now we do not have any BTS but we are going to have nano BTS it soon. So we are trying to communicate between BSC and MSC only without having BTS.
We created openbsc.cfg file and tried to avoid the BTS configuration in it and run but with no success. So, I would appreciate your help.
On Thu, Oct 7, 2010 at 12:18 PM, Holger Hans Peter Freyther < holger@freyther.de> wrote:
On 10/07/2010 05:12 PM, tejas oza wrote:
Hi Holger,
We will publish our code as soon as we produced it.
As you said we made C node and wrapped libsccp.a. After running
bsc_msc_ip we
are getting following error.
*ozmank@ozmank-desktop:~/libosmocore/openbsc/openbsc/src$ ./bsc_msc_ip 130.238.15.219 58620 Bootstraping the network. Sending GSM08.08 reset. mi_e1_line_update could not open socket Address family not supported by
protocol
full talloc report on 'sccp' (total 29 bytes in 2 blocks) struct sccp_data_callback contains 28 bytes in 1 blocks
(ref 0)
0x9ca12a8*
Can you please advice.
Well,
what do you think that this error could be? Which kind of BTS do you attempt to use? OpenBSC supports the Siemens BS11 and the ipaccess nanoBTS, with a configuration file (that is documented in the wiki) you define which BTS to use... you have obviously created a openbsc.cfg otherwise it would have complained about a missing one.
On 10/07/2010 08:45 PM, tejas oza wrote:
Hi Holger,
Right now we do not have any BTS but we are going to have nano BTS it soon. So we are trying to communicate between BSC and MSC only without having BTS.
We created openbsc.cfg file and tried to avoid the BTS configuration in it and run but with no success.
Well, you decided to copy to a config for the Siemens BS11... if you copy one of the examples for the nanoBTS the BSC should come up, if your address of the MSC (nowadays it is better to put it into the bsc.cfg) is correct, the BSC will connect to it and send a GSM0808 RESET to it.
All you can test without a BSC is to send Paging messages from the MSC to the BSC (as they do not require any connection).
Do you have a link to your university project?
Hi Holger,
We are able to connect to MSC but after it gets connected, it throws below error:-
ozmank@ozmank-desktop:~/libosmocore/openbsc/openbsc/src$ ./bsc_msc_ipBootstraping the network. Sending GSM08.08 reset. <000f> bsc_msc.c:135 Attempting to connect MSC at 130.238.15.219 :58620 <000f> bsc_msc.c:168 MSC Connection in progress <000f> bsc_msc.c:97 (Re)Connected to the MSC. <000d> input/ipaccess.c:323 Can not read this packet. 12336 avail Failed to parse ip access message: -5 <000d> input/ipaccess.c:331 short read! Got 4 from 114 Failed to parse ip access message: -5 <000f> bsc_msc_ip.c:1047 The connection to the MSC was lost. <000f> bsc_msc_ip.c:976 Lost MSC connection. Freing stuff.
So We are wonder that if we are sending some wrong packet OR wrong packet format from MSC to BSC as it is throwing cannot read the packet.
Can you Please advice.
On Thu, Oct 7, 2010 at 3:02 PM, Holger Hans Peter Freyther < holger@freyther.de> wrote:
On 10/07/2010 08:45 PM, tejas oza wrote:
Hi Holger,
Right now we do not have any BTS but we are going to have nano BTS it
soon.
So we are trying to communicate between BSC and MSC only without having
BTS.
We created openbsc.cfg file and tried to avoid the BTS configuration in
it and
run but with no success.
Well, you decided to copy to a config for the Siemens BS11... if you copy one of the examples for the nanoBTS the BSC should come up, if your address of the MSC (nowadays it is better to put it into the bsc.cfg) is correct, the BSC will connect to it and send a GSM0808 RESET to it.
All you can test without a BSC is to send Paging messages from the MSC to the BSC (as they do not require any connection).
Do you have a link to your university project?
Hi Tejas,
On Mon, Oct 11, 2010 at 04:13:42PM +0200, tejas oza wrote:
We are able to connect to MSC but after it gets connected, it throws below error:- [...] So We are wonder that if we are sending some wrong packet OR wrong packet format from MSC to BSC as it is throwing cannot read the packet.
Can you Please advice.
How can we give any advice, if you don't include packet traces (.pcap file) of the actual communication? Only if we have this information, we can make any assessment at all..
On 10/12/2010 01:15 PM, Harald Welte wrote:
Can you Please advice.
How can we give any advice, if you don't include packet traces (.pcap file) of the actual communication? Only if we have this information, we can make any assessment at all..
You can download the latest version of wireshark, it will detect the IPA protocol on port 5000. Two things to note on the stream..
a) use the right header.. e.g. the length (two byte) is in network order b) send header + payload together... our code makes that assumption
Hi Holger,
We managed to send the paging request from MSC to OpenBSC and we can track it on wireshark that the packet format is correct but we are wonder if we should expect any response from OpenBSC to MSC for paging request and OpenBSC code seems to be returning -1 as per our understanding.
Kindly find the attachment for .pcap file.
On Wed, Oct 13, 2010 at 1:04 AM, Holger Hans Peter Freyther < holger@freyther.de> wrote:
On 10/12/2010 01:15 PM, Harald Welte wrote:
Can you Please advice.
How can we give any advice, if you don't include packet traces (.pcap
file)
of the actual communication? Only if we have this information, we can
make
any assessment at all..
You can download the latest version of wireshark, it will detect the IPA protocol on port 5000. Two things to note on the stream..
a) use the right header.. e.g. the length (two byte) is in network order b) send header + payload together... our code makes that assumption
On 10/14/2010 10:08 AM, tejas oza wrote:
Hi Holger,
We managed to send the paging request from MSC to OpenBSC and we can track it on wireshark that the packet format is correct but we are wonder if we should expect any response from OpenBSC to MSC for paging request and OpenBSC code seems to be returning -1 as per our understanding.
Theory: After you sent the Paging Request the BSC will start to page the subscriber. If the Mobile Stations sees the Paging Request it will open a channel with the BSC, the BSC will forward the Paging Response to the MSC on a new SCCP connection. Ironically the code you are using has a bsc_msc_ip.c:test_msc that is called when you send the ID_GET request. This will send you the data of a paging response, so you could start dealing with connections while we figure out the current issue.
Issue: You are right the package looks good. You will need to go to the code to figure out what is wrong. You should set a breakpoint in ipaccess_read_msg (src/input/ipaccess.c). It will try to receive the header and then it will receive the data. In the other mail I have indicated that you need to send header + data as one packet (instead of two writes for header and data).
You can use gdb to step through these lines of code and you should see the place where it returns NULL.
Hi Holger,
We have wrapped the libsccp.a to our C node so can you let us know how it can be useful to write C Node as we run the code without wrapping the libsccp.a and it works fine.
We mean how can we make the best use of it !
On Thu, Oct 14, 2010 at 10:24 AM, Holger Hans Peter Freyther < holger@freyther.de> wrote:
On 10/14/2010 10:08 AM, tejas oza wrote:
Hi Holger,
We managed to send the paging request from MSC to OpenBSC and we can
track it
on wireshark that the packet format is correct but we are wonder if we
should
expect any response from OpenBSC to MSC for paging request and OpenBSC
code
seems to be returning -1 as per our understanding.
Theory: After you sent the Paging Request the BSC will start to page the subscriber. If the Mobile Stations sees the Paging Request it will open a channel with the BSC, the BSC will forward the Paging Response to the MSC on a new SCCP connection. Ironically the code you are using has a bsc_msc_ip.c:test_msc that is called when you send the ID_GET request. This will send you the data of a paging response, so you could start dealing with connections while we figure out the current issue.
Issue: You are right the package looks good. You will need to go to the code to figure out what is wrong. You should set a breakpoint in ipaccess_read_msg (src/input/ipaccess.c). It will try to receive the header and then it will receive the data. In the other mail I have indicated that you need to send header + data as one packet (instead of two writes for header and data).
You can use gdb to step through these lines of code and you should see the place where it returns NULL.
On 10/15/2010 10:15 AM, tejas oza wrote:
Hi Holger,
We have wrapped the libsccp.a to our C node so can you let us know how it can be useful to write C Node as we run the code without wrapping the libsccp.a and it works fine.
Hi,
did you resolve the other problem? I am not sure I understand your current question?
No not yet. We are working on that. And regarding the question we had some confusion on libsccp.a that how can we know which functions are defined in this library so that we can use them.
On Fri, Oct 15, 2010 at 12:14 PM, Holger Hans Peter Freyther < holger@freyther.de> wrote:
On 10/15/2010 10:15 AM, tejas oza wrote:
Hi Holger,
We have wrapped the libsccp.a to our C node so can you let us know how it
can
be useful to write C Node as we run the code without wrapping the
libsccp.a
and it works fine.
Hi,
did you resolve the other problem? I am not sure I understand your current question?
On 10/15/2010 05:16 PM, tejas oza wrote:
No not yet. We are working on that. And regarding the question we had some confusion on libsccp.a that how can we know which functions are defined in this library so that we can use them.
well check the header file called sccp.h? It contains functions to handle the connections, but also functions to just parse and generate packets.
On 10/15/2010 05:27 PM, Holger Hans Peter Freyther wrote:
well check the header file called sccp.h? It contains functions to handle the connections, but also functions to just parse and generate packets.
The C Node that I implemented can be used to unpack/pack UDT messages. You could do something similar for Connection Oriented SCCP Messages. Otherwise you will need to generate and parse SCCP from within Erlang.
[1] http://git.osmocom.org/gitweb?p=osmo-map-masq.git;a=blob;f=c_src/sccp_node.c...
Hi Holger,
We found the below code in bsc_msc_ip.c
static int msc_sccp_accept(struct sccp_connection *connection, void *data) { LOGP(DMSC, LOGL_DEBUG, "Rejecting incoming SCCP connection.\n"); return -1; }
So, Does the Open BSC accepts the SCCP connection request from MSC ?
On Fri, Oct 15, 2010 at 5:30 PM, Holger Hans Peter Freyther < holger@freyther.de> wrote:
On 10/15/2010 05:27 PM, Holger Hans Peter Freyther wrote:
well check the header file called sccp.h? It contains functions to handle
the
connections, but also functions to just parse and generate packets.
The C Node that I implemented can be used to unpack/pack UDT messages. You could do something similar for Connection Oriented SCCP Messages. Otherwise you will need to generate and parse SCCP from within Erlang.
[1]
http://git.osmocom.org/gitweb?p=osmo-map-masq.git;a=blob;f=c_src/sccp_node.c...
On 10/26/2010 11:10 AM, tejas oza wrote:
Hi Holger,
We found the below code in bsc_msc_ip.c
static int msc_sccp_accept(struct sccp_connection *connection, void *data) { LOGP(DMSC, LOGL_DEBUG, "Rejecting incoming SCCP connection.\n"); return -1; }
So, Does the Open BSC accepts the SCCP connection request from MSC ?
As you have found in the code. the BSC does not accept connections from the MSC. The reason is that I have not found anything in GSM08.08 that indicates that the MSC will open a connection. Am I wrong with that? Or is your new MSC trying to open a connection?
i hope this helps z.
PS: According to my understanding of the netiqutte you would have to create a new top level post for that. PPS: Do you have a time plan on contributing the POI fixlet?
Hi Holger,
Kindly find the Attachment for SCCP Patch.
On Tue, Oct 26, 2010 at 12:12 PM, Holger Hans Peter Freyther < holger@freyther.de> wrote:
On 10/26/2010 11:10 AM, tejas oza wrote:
Hi Holger,
We found the below code in bsc_msc_ip.c
static int msc_sccp_accept(struct sccp_connection *connection, void
*data)
{ LOGP(DMSC, LOGL_DEBUG, "Rejecting incoming SCCP connection.\n"); return -1; }
So, Does the Open BSC accepts the SCCP connection request from MSC ?
As you have found in the code. the BSC does not accept connections from the MSC. The reason is that I have not found anything in GSM08.08 that indicates that the MSC will open a connection. Am I wrong with that? Or is your new MSC trying to open a connection?
i hope this helps z.
PS: According to my understanding of the netiqutte you would have to create a new top level post for that. PPS: Do you have a time plan on contributing the POI fixlet?
On 10/26/2010 12:39 PM, tejas oza wrote:
Hi Holger,
Kindly find the Attachment for SCCP Patch.
ah, I had described something else. E.g. there is one other place where the offsets are wrong. This is why I hinted that create_sccp_addr should return the bytes it consumed.
The benefit is that we do not need to add a lot of "if (addr->this)" to the code that is calling create_sccp_addr.
I am going to prepare a patch like this.
Hi zecke,
getting back t oan old thread:
On Tue, Oct 26, 2010 at 12:12:12PM +0200, Holger Hans Peter Freyther wrote:
On 10/26/2010 11:10 AM, tejas oza wrote:
Hi Holger,
We found the below code in bsc_msc_ip.c
static int msc_sccp_accept(struct sccp_connection *connection, void *data) { LOGP(DMSC, LOGL_DEBUG, "Rejecting incoming SCCP connection.\n"); return -1; }
So, Does the Open BSC accepts the SCCP connection request from MSC ?
As you have found in the code. the BSC does not accept connections from the MSC. The reason is that I have not found anything in GSM08.08 that indicates that the MSC will open a connection. Am I wrong with that?
GSM 08.06 Chapter 6.1 Connection establishment: ii) the MSC decides to perform an external handover and a new dedicated radio resource has t obe reserved in the new BSS. The SCCP connection establishment is then initiated by the MSC
So there is at least one case in which this procedure is being used. I know, we only do intra-BSC handover at the moment and no external handover is supported.
But I think you might be interested in this anyway :)
Regards, Harald
On 12/14/2010 10:36 AM, Harald Welte wrote:
Hi zecke,
GSM 08.06 Chapter 6.1 Connection establishment: ii) the MSC decides to perform an external handover and a new dedicated radio resource has t obe reserved in the new BSS. The SCCP connection establishment is then initiated by the MSC
So there is at least one case in which this procedure is being used. I know, we only do intra-BSC handover at the moment and no external handover is supported.
But I think you might be interested in this anyway :)
Ah okay, I will need to look at 08.06 as well. Thanks a lot for finding it. And yes we do not support inter-BSC hand-over and even intra-BSC handover is slightly wrong ad we do not inform the MSC about the handover (leaving alone the mgcp code to handle that).
holger
-
Harald Welte -
Holger Hans Peter Freyther -
Stefan Schmidt -
tejas oza