- OpenBSC - lists.osmocom.org

Questions regarding sysmoBTS2050 power management
by Holger Hans Peter Freyther 23 Jun '14

23 Jun '14

hi! I would like to have your opinion on the power management of the sysmoBTS2050 hardware in various conditions: Using only a single TRX with 10: We have to disable the second TRX. Would you want to set a different device id in the EEPROM for the sysmoBTS2050 to indicate a single TRX board? If yes we could have nominal power return 40 and something like the sysmobts-mgr could disable the power of the second trx on start? Another option would be to make it configurable inside the bts configuration file. In this case the bts process would need to tell the sysmobts-mgr to switch off the second trx? Reducing power in a dual-bts setup: The current unfinished idea would be that in case the system heats up too much we reduce the transmit power on the first TRX. Is this enough? Is this enough because in the long run we will manage both TRX from the first one? Or shall we handle the heat inside OpenBSC to temporarily reduce the power? This way we would need to send information in case the system returns to an acceptable temperature? holger

2 1

Bounce action notification
by mailman＠lists.osmocom.org 22 Jun '14

22 Jun '14

This is a Mailman mailing list bounce action notice: List: OpenBSC Member: mki(a)mki-consult.de Action: Subscription disabled. Reason: Excessive or fatal bounces. The triggering bounce notice is attached below. Questions? Contact the Mailman site administrator at mailman(a)lists.osmocom.org.

1 0

Patch: handover check for E1 based BTS
by Sipos Csaba 18 Jun '14

18 Jun '14

The function that checks for HO compatibility is wrong for quite some time. Even for E1 based BTSs the function requires RTP proxy to be enabled. The following patch is fixing this, by changing the condition so RTP proxy is only required, when the BTS is an IP based unit: diff --git a/openbsc/src/libbsc/bsc_vty.c b/openbsc/src/libbsc/bsc_vty.c index 6acf0c6..fbf28ec 100644 --- a/openbsc/src/libbsc/bsc_vty.c +++ b/openbsc/src/libbsc/bsc_vty.c @@ -1427,8 +1427,9 @@ DEFUN(cfg_net_handover, cfg_net_handover_cmd, { int enable = atoi(argv[0]); struct gsm_network *gsmnet = gsmnet_from_vty(vty); + struct gsm_bts *bts = vty->index; - if (enable && ipacc_rtp_direct) { + if (enable && ((is_ipaccess_bts(bts)) && ipacc_rtp_direct)) { vty_out(vty, "%% Cannot enable handover unless RTP Proxy mode " "is enabled by using the -P command line option%s", VTY_NEWLINE); The patch is tested with Nokia Insite. I hope someone can commit this to the master branch. Regards, Csaba

1 0

Unable to build openbsc due to missing 'libosmo-netif' package
by hulds＠gmx.net 17 Jun '14

17 Jun '14

Hi, i try to get openbsc to work. To do that i followed the steps on the page : http://openbsc.osmocom.org/trac/wiki/OpenBSC_GPRS till the last step everything go's ok, during the last step : cd /root/openbsc/openbsc; autoreconf -fi; export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig; ./configure; make I get the message : checking for LIBOSMONETIF... no configure: error: Package requirements (libosmo-netif >= 0.0.1) were not met: No package 'libosmo-netif' found I tried to find this package but was not able to find it. Do you know where i can find this package so that i can get things working ? Regards Henry

2 1

[PATCH] mgcp/test: Add test case using the fmtp_extra info
by Jacob Erlbeck 15 Jun '14

15 Jun '14

Add tests setting the fmtp_extra field to check the response generation. This triggers a bug found by Coverity. Addresses: Coverity CID 1220873 Sponsored-by: On-Waves ehf --- openbsc/tests/mgcp/mgcp_test.c | 31 +++++++++++++++++++++++++++++++ openbsc/tests/mgcp/mgcp_test.ok | 38 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 69 insertions(+) diff --git a/openbsc/tests/mgcp/mgcp_test.c b/openbsc/tests/mgcp/mgcp_test.c index 19615d9..e185fbc 100644 --- a/openbsc/tests/mgcp/mgcp_test.c +++ b/openbsc/tests/mgcp/mgcp_test.c @@ -20,6 +20,7 @@ #define _GNU_SOURCE #include <openbsc/mgcp.h> +#include <openbsc/vty.h> #include <openbsc/mgcp_internal.h> #include <osmocom/core/application.h> @@ -88,6 +89,17 @@ static void test_strline(void) "m=audio 0 RTP/AVP 126\r\n" \ "a=rtpmap:126 AMR/8000\r\n" \ "a=ptime:20\r\n" +#define MDCX3_FMTP_RET "200 18983215 OK\r\n" \ + "I: 3\n" \ + "\n" \ + "v=0\r\n" \ + "o=- 3 23 IN IP4 0.0.0.0\r\n" \ + "c=IN IP4 0.0.0.0\r\n" \ + "t=0 0\r\n" \ + "m=audio 0 RTP/AVP 126\r\n" \ + "a=rtpmap:126 AMR/8000\r\n" \ + "a=fmtp:126 0/1/2\r\n" \ + "a=ptime:20\r\n" #define MDCX4 "MDCX 18983216 1@mgw MGCP 1.0\r\n" \ "M: sendrecv\r" \ "C: 2\r\n" \ @@ -203,6 +215,18 @@ static void test_strline(void) "a=rtpmap:126 AMR/8000\r\n" \ "a=ptime:20\r\n" +#define CRCX_FMTP_RET "200 2 OK\r\n" \ + "I: 3\n" \ + "\n" \ + "v=0\r\n" \ + "o=- 3 23 IN IP4 0.0.0.0\r\n" \ + "c=IN IP4 0.0.0.0\r\n" \ + "t=0 0\r\n" \ + "m=audio 0 RTP/AVP 126\r\n" \ + "a=rtpmap:126 AMR/8000\r\n" \ + "a=fmtp:126 0/1/2\r\n" \ + "a=ptime:20\r\n" + #define CRCX_ZYN "CRCX 2 1@mgw MGCP 1.0\r" \ "M: recvonly\r" \ "C: 2\r\r" \ @@ -250,6 +274,8 @@ struct mgcp_test { const char *exp_resp; int exp_net_ptype; int exp_bts_ptype; + + char *extra_fmtp; }; static const struct mgcp_test tests[] = { @@ -275,6 +301,9 @@ static const struct mgcp_test tests[] = { { "RQNT1", RQNT, RQNT1_RET }, { "RQNT2", RQNT2, RQNT2_RET }, { "DLCX", DLCX, DLCX_RET, -1, -1 }, + { "CRCX", CRCX, CRCX_FMTP_RET, 97, 126, .extra_fmtp = "a=fmtp:126 0/1/2" }, + { "MDCX3", MDCX3, MDCX3_FMTP_RET, PTYPE_NONE, 126 , .extra_fmtp = "a=fmtp:126 0/1/2" }, + { "DLCX", DLCX, DLCX_RET, -1, -1 , .extra_fmtp = "a=fmtp:126 0/1/2" }, }; static const struct mgcp_test retransmit[] = { @@ -398,6 +427,8 @@ static void test_messages(void) last_endpoint = -1; dummy_packets = 0; + bsc_replace_string(cfg, &cfg->trunk.audio_fmtp_extra, t->extra_fmtp); + inp = create_msg(t->req); msg = mgcp_handle_message(cfg, inp); msgb_free(inp); diff --git a/openbsc/tests/mgcp/mgcp_test.ok b/openbsc/tests/mgcp/mgcp_test.ok index 7301a81..6c5a6bc 100644 --- a/openbsc/tests/mgcp/mgcp_test.ok +++ b/openbsc/tests/mgcp/mgcp_test.ok @@ -75,6 +75,44 @@ Detected packet duration: 20 Requested packetization period not set Connection mode: 0: NONE Testing CRCX +CRCX failed '200 2 OK +I: 3 + +v=0 +o=- 3 23 IN IP4 0.0.0.0 +c=IN IP4 0.0.0.0 +t=0 0 +m=audio 0 RTP/AVP 126 +a=rtpmap:126 AMR/8000 +a=rtpmap:126 AMR/8000 +a=ptime:20 +' +Dummy packets: 1 +Detected packet duration: 40 +Requested packetetization period: 20-20 +Connection mode: 1: RECV +Testing MDCX3 +MDCX3 failed '200 18983215 OK +I: 3 + +v=0 +o=- 3 23 IN IP4 0.0.0.0 +c=IN IP4 0.0.0.0 +t=0 0 +m=audio 0 RTP/AVP 126 +a=rtpmap:126 AMR/8000 +a=rtpmap:126 AMR/8000 +a=ptime:20 +' +Dummy packets: 1 +Packet duration not set +Requested packetization period not set +Connection mode not set +Testing DLCX +Detected packet duration: 20 +Requested packetization period not set +Connection mode: 0: NONE +Testing CRCX Re-transmitting CRCX Testing RQNT1 Re-transmitting RQNT1 -- 1.9.1

2 1

XID parameters
by Josh toal 13 Jun '14

13 Jun '14

Hi All As per the specs after PDP context accept mobile station should exchange XID parameters with SGSN. before it initiate transfer of data. As per the literature mobile station if not happy with the negotiated quality of service it will send the deactivation request and I am getting the same. Is this is the only reason mobile station will send deactivation request or something else, which I am missing. Any suggestion to fix this problem. can anybody comment on this please. regards Josh

2 2

[PATCH] mgcp: Fix SDP formatting of fmtp_extra (Coverity)
by Jacob Erlbeck 12 Jun '14

12 Jun '14

Currently when ftmp_extra is set, a doubled a=rtpmap line is emitted instead of the fmtp_extra info. This patch fixes replaces the formerly copied and pasted but not modified snprintf parameters by the correct ones. Fixes: Coverity CID 1220873 Sponsored-by: On-Waves ehf --- openbsc/src/libmgcp/mgcp_protocol.c | 3 +-- openbsc/tests/mgcp/mgcp_test.ok | 24 ------------------------ 2 files changed, 1 insertion(+), 26 deletions(-) diff --git a/openbsc/src/libmgcp/mgcp_protocol.c b/openbsc/src/libmgcp/mgcp_protocol.c index f0457d1..6e48949 100644 --- a/openbsc/src/libmgcp/mgcp_protocol.c +++ b/openbsc/src/libmgcp/mgcp_protocol.c @@ -287,8 +287,7 @@ static int write_response_sdp(struct mgcp_endpoint *endp, if (fmtp_extra) { nchars = snprintf(sdp_record + len, size - len, - "a=rtpmap:%d %s\r\n", - payload_type, audio_name); + "%s\r\n", fmtp_extra); if (nchars < 0 || nchars >= size - len) goto buffer_too_small; diff --git a/openbsc/tests/mgcp/mgcp_test.ok b/openbsc/tests/mgcp/mgcp_test.ok index 6c5a6bc..033f783 100644 --- a/openbsc/tests/mgcp/mgcp_test.ok +++ b/openbsc/tests/mgcp/mgcp_test.ok @@ -75,35 +75,11 @@ Detected packet duration: 20 Requested packetization period not set Connection mode: 0: NONE Testing CRCX -CRCX failed '200 2 OK -I: 3 - -v=0 -o=- 3 23 IN IP4 0.0.0.0 -c=IN IP4 0.0.0.0 -t=0 0 -m=audio 0 RTP/AVP 126 -a=rtpmap:126 AMR/8000 -a=rtpmap:126 AMR/8000 -a=ptime:20 -' Dummy packets: 1 Detected packet duration: 40 Requested packetetization period: 20-20 Connection mode: 1: RECV Testing MDCX3 -MDCX3 failed '200 18983215 OK -I: 3 - -v=0 -o=- 3 23 IN IP4 0.0.0.0 -c=IN IP4 0.0.0.0 -t=0 0 -m=audio 0 RTP/AVP 126 -a=rtpmap:126 AMR/8000 -a=rtpmap:126 AMR/8000 -a=ptime:20 -' Dummy packets: 1 Packet duration not set Requested packetization period not set -- 1.9.1

1 0

[PATCH 1/2] write_queue: Avoid possible use-after-free if fd is read-/writable
by Daniel Willmann 12 Jun '14

12 Jun '14

If the FD is both readable and writable and the read callback closes the connection (and frees the surrounding structure) we shouldn't call the write callback (or check anything else in the read fd). With this patch callback functions can return -EBADFD if they don't want the FD to be handled any more. --- src/write_queue.c | 27 ++++++++++++++++++++++----- 1 file changed, 22 insertions(+), 5 deletions(-) diff --git a/src/write_queue.c b/src/write_queue.c index cef40f8..d17493d 100644 --- a/src/write_queue.c +++ b/src/write_queue.c @@ -21,8 +21,15 @@ * */ +#include <errno.h> #include <osmocom/core/write_queue.h> +#define HANDLE_BAD_FD(rc, label) \ + do { \ + if (rc == -EBADFD) \ + goto label; \ + } while (0); + /*! \addtogroup write_queue * @{ */ @@ -39,14 +46,19 @@ int osmo_wqueue_bfd_cb(struct osmo_fd *fd, unsigned int what) { struct osmo_wqueue *queue; + int rc; queue = container_of(fd, struct osmo_wqueue, bfd); - if (what & BSC_FD_READ) - queue->read_cb(fd); + if (what & BSC_FD_READ) { + rc = queue->read_cb(fd); + HANDLE_BAD_FD(rc, err_badfd); + } - if (what & BSC_FD_EXCEPT) - queue->except_cb(fd); + if (what & BSC_FD_EXCEPT) { + rc = queue->except_cb(fd); + HANDLE_BAD_FD(rc, err_badfd); + } if (what & BSC_FD_WRITE) { struct msgb *msg; @@ -58,16 +70,21 @@ int osmo_wqueue_bfd_cb(struct osmo_fd *fd, unsigned int what) --queue->current_length; msg = msgb_dequeue(&queue->msg_queue); - queue->write_cb(fd, msg); + rc = queue->write_cb(fd, msg); msgb_free(msg); + HANDLE_BAD_FD(rc, err_badfd); + if (!llist_empty(&queue->msg_queue)) fd->when |= BSC_FD_WRITE; } } return 0; +err_badfd: + return rc; } +#undef HANDLE_BAD_FD /*! \brief Initialize a \ref osmo_wqueue structure * \param[in] queue Write queue to operate on -- 1.8.4.2

4 6

[osmo-pcu 1/1] bts: Remove outdated comment
by Daniel Willmann 04 Jun '14

04 Jun '14

There is no parameter in the function and the bts has a memeber trx which has a member pdch. Sponsored-by: On-Waves ehf --- src/bts.cpp | 1 - 1 file changed, 1 deletion(-) diff --git a/src/bts.cpp b/src/bts.cpp index 9de6dc3..3dd5705 100644 --- a/src/bts.cpp +++ b/src/bts.cpp @@ -532,7 +532,6 @@ void gprs_rlcmac_pdch::disable() m_is_enabled = 0; } -/* TODO: kill the parameter and make a pdch belong to a trx.. to a bts.. */ void gprs_rlcmac_pdch::free_resources() { struct gprs_rlcmac_paging *pag; -- 1.8.4.2

1 0

[osmo-pcu 1/1] Always exit and don't try to recover
by Daniel Willmann 04 Jun '14

04 Jun '14

From: Daniel Willmann <dwillmann(a)sysmocom.de> The current code tries to recover from dropped connections and resets the pcu state so it can keep running. However, this never worked correctly which is why the -e option is used. This option exits the pcu as soon as the internal state needs to be reset. This patch removes this option and makes this behaviour default. Ticket: SYS#390 Sponsored-by: On-Waves ehf --- src/gprs_bssgp_pcu.cpp | 13 +------------ src/gprs_bssgp_pcu.h | 4 +--- src/openbts_sock.cpp | 4 +++- src/pcu_l1_if.cpp | 4 ++-- src/pcu_main.cpp | 3 +-- src/sysmo_sock.cpp | 8 ++------ 6 files changed, 10 insertions(+), 26 deletions(-) diff --git a/src/gprs_bssgp_pcu.cpp b/src/gprs_bssgp_pcu.cpp index 133eb97..6c5d72c 100644 --- a/src/gprs_bssgp_pcu.cpp +++ b/src/gprs_bssgp_pcu.cpp @@ -533,13 +533,8 @@ struct gprs_bssgp_pcu *gprs_bssgp_create_and_connect(struct gprs_rlcmac_bts *bts return &the_pcu; } -void gprs_bssgp_destroy_or_exit(void) +void gprs_bssgp_destroy(void) { - if (the_pcu.exit_on_destroy) { - LOGP(DBSSGP, LOGL_NOTICE, "Exiting on BSSGP destruction.\n"); - exit(0); - } - if (!bssgp_nsi) return; @@ -564,12 +559,6 @@ void gprs_bssgp_destroy_or_exit(void) bssgp_nsi = NULL; } -void gprs_bssgp_exit_on_destroy(void) -{ - LOGP(DBSSGP, LOGL_NOTICE, "Going to quit on BSSGP destruction\n"); - the_pcu.exit_on_destroy = 1; -} - struct bssgp_bvc_ctx *gprs_bssgp_pcu_current_bctx(void) { return the_pcu.bctx; diff --git a/src/gprs_bssgp_pcu.h b/src/gprs_bssgp_pcu.h index aedcfc0..32b6728 100644 --- a/src/gprs_bssgp_pcu.h +++ b/src/gprs_bssgp_pcu.h @@ -57,7 +57,6 @@ struct gprs_bssgp_pcu { int bvc_sig_reset; int bvc_reset; int bvc_unblocked; - int exit_on_destroy; /** callbacks below */ @@ -75,8 +74,7 @@ struct gprs_bssgp_pcu *gprs_bssgp_create_and_connect(struct gprs_rlcmac_bts *bts uint16_t nsvci, uint16_t bvci, uint16_t mcc, uint16_t mnc, uint16_t lac, uint16_t rac, uint16_t cell_id); -void gprs_bssgp_exit_on_destroy(void); -void gprs_bssgp_destroy_or_exit(void); +void gprs_bssgp_destroy(void); struct bssgp_bvc_ctx *gprs_bssgp_pcu_current_bctx(void); diff --git a/src/openbts_sock.cpp b/src/openbts_sock.cpp index a09f834..2d9cae4 100644 --- a/src/openbts_sock.cpp +++ b/src/openbts_sock.cpp @@ -179,8 +179,10 @@ int pcu_l1if_open() void pcu_l1if_close(void) { - gprs_bssgp_destroy_or_exit(); + gprs_bssgp_destroy(); /* FIXME: cleanup l1if */ talloc_free(l1fh->fl1h); + + exit(0); } diff --git a/src/pcu_l1_if.cpp b/src/pcu_l1_if.cpp index ce13b96..0cb79eb 100644 --- a/src/pcu_l1_if.cpp +++ b/src/pcu_l1_if.cpp @@ -324,8 +324,8 @@ bssgp_failed: for (ts = 0; ts < 8; ts++) bts->trx[trx].pdch[ts].free_resources(); } - gprs_bssgp_destroy_or_exit(); - return 0; + gprs_bssgp_destroy(); + exit(0); } LOGP(DL1IF, LOGL_INFO, "BTS available\n"); LOGP(DL1IF, LOGL_DEBUG, " mcc=%x\n", info_ind->mcc); diff --git a/src/pcu_main.cpp b/src/pcu_main.cpp index c0dac14..8631ad3 100644 --- a/src/pcu_main.cpp +++ b/src/pcu_main.cpp @@ -56,7 +56,6 @@ static void print_help() "provided by BTS\n" " -r --realtime PRIO Use SCHED_RR with the specified " "priority\n" - " -e --exit Exit the application on disconnect\n" ); } @@ -105,7 +104,7 @@ static void handle_options(int argc, char **argv) rt_prio = atoi(optarg); break; case 'e': - gprs_bssgp_exit_on_destroy(); + fprintf(stderr, "Warning: Option '-e' is deprecated!\n"); break; default: fprintf(stderr, "Unknown option '%c'\n", c); diff --git a/src/sysmo_sock.cpp b/src/sysmo_sock.cpp index 2e2d9d3..951653e 100644 --- a/src/sysmo_sock.cpp +++ b/src/sysmo_sock.cpp @@ -112,12 +112,8 @@ static void pcu_sock_close(struct pcu_sock_state *state, int lost) gprs_rlcmac_tbf::free_all(&bts->trx[trx]); } - gprs_bssgp_destroy_or_exit(); - - if (lost) { - state->timer.cb = pcu_sock_timeout; - osmo_timer_schedule(&state->timer, 5, 0); - } + gprs_bssgp_destroy(); + exit(0); } static int pcu_sock_read(struct osmo_fd *bfd) -- 1.8.4.2

1 0