sr-dev October 2009

sr-dev@lists.kamailio.org

30 participants
468 discussions

pseudo variable $ad is not working
by jh＠tutpro.com 28 Oct '09

28 Oct '09

i have this kind of test in my config: if (lookup_domain("$fd", "$ad")) { $var(lcr_id) = $fd.lcr_id; } else { xlog("L_WARN", "Domain $ad of authorized user $aU is not local!\n"); send_reply("404", "Not found"); exit; }; the problem is that when request that has header Proxy-Authorization: Digest username="test(a)test.fi" arrives, sr prints to syslog Oct 28 20:01:09 localhost /usr/sbin/sip-proxy[7248]: WARNING: Domain of authorized user test is not local! i.e., $ad does not have test.fi as its value. according to wiki it should work: Auth username domain $ad - domain part of username from Authorization or Proxy-Authorization header so this looks like a bug to me. -- juha

1 1

git:master: Merge remote branch 'origin/sr_3.0'
by Andrei Pelinescu-Onciul 28 Oct '09

28 Oct '09

Module: sip-router Branch: master Commit: 8b9ccd17e24054d734c12be2c2ae79bc864bf59f URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=8b9ccd1… Author: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Committer: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Date: Wed Oct 28 16:41:39 2009 +0100 Merge remote branch 'origin/sr_3.0' * origin/sr_3.0: htable(k): fix non-init act. ctx in event route execution tm: fix/support changing r-uris and path in branch routes tm: support for changing dst_uri in branch routes registrar: Fix handling of cases where contacts > max_contacts domain: Do not report errors when domain cannot be extracted from URI. parse_sip_msg_uri: Log broken URIs only when debugging is enabled. modules/lcr: documentation improvement sctp: count rejects sent to the remote peer (stats) avp_db: Removes a spurious error message. ctl: missing ifdef (minor) event parser: Add missing string boundary checks to event_parser func. tm: Number of fixes in code and documentation for serial forking. Implements function reset_path_vector. ---

1 0

git:sr_3.0: tm: fix/support changing r-uris and path in branch routes
by Andrei Pelinescu-Onciul 28 Oct '09

28 Oct '09

Module: sip-router Branch: sr_3.0 Commit: 610cb930bee937c744468c6781aeed8b17b4b7d0 URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=610cb93… Author: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Committer: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Date: Wed Oct 28 15:50:45 2009 +0100 tm: fix/support changing r-uris and path in branch routes - changing the r-uri in a branch route is now safe and the change will propagate to branches created due to dns failover. - changing path in a branch route is now allowed and it will propagate to possible dns failover branches. - optimize the no branch route & no callback case. - build the branch buffer directly into shm mem. - s/prepare_uac_request/prepare_new_uac/ - prepare_new_uac() takes care now also of setting the branch retr. buffer, uri and path (functionality moved from add_uac()) --- modules/tm/t_fwd.c | 417 +++++++++++++++++++++++++++++++--------------------- 1 files changed, 247 insertions(+), 170 deletions(-) Diff: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commitdiff;h=610…

1 1

git:sr_3.0: htable(k): fix non-init act. ctx in event route execution
by Andrei Pelinescu-Onciul 28 Oct '09

28 Oct '09

Module: sip-router Branch: sr_3.0 Commit: 4ac8656cb483f64e88a2a70996178ffdef85e097 URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=4ac8656… Author: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Committer: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Date: Wed Oct 28 16:11:22 2009 +0100 htable(k): fix non-init act. ctx in event route execution - the event route was executed with an uninitialized run_act_ctx. --- modules_k/htable/htable.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/modules_k/htable/htable.c b/modules_k/htable/htable.c index 6291c2b..2d49864 100644 --- a/modules_k/htable/htable.c +++ b/modules_k/htable/htable.c @@ -185,6 +185,7 @@ static int child_init(int rank) fmsg = faked_msg_next(); rtb = get_route_type(); set_route_type(REQUEST_ROUTE); + init_run_actions_ctx(&ctx); run_top_route(event_rt.rlist[rt], fmsg, &ctx); if(ctx.run_flags&DROP_R_F) {

1 0

git:sr_3.0: tm: support for changing dst_uri in branch routes
by Andrei Pelinescu-Onciul 28 Oct '09

28 Oct '09

Module: sip-router Branch: sr_3.0 Commit: 44f7bf621476b18e5df5e45161305a3bce22d498 URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=44f7bf6… Author: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Committer: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Date: Wed Oct 28 08:36:47 2009 +0100 tm: support for changing dst_uri in branch routes - dst_uri can now be changed in branch routes, if the branch route is not called during dns failover (which never happens, but might be enabled in the future). - dst_uri can now be checked in branch routes. The dst_uri on branch route entry will be the branch next hop (initial next hop and not the dns resolved one) if a t_relay() based on uri was used, or empty if a destination was forced (e.g. t_relay_to_udp(127.0.0.1, 9)). - s/print_uac_request/prepare_uac_request to better reflect what is actually doing - prepare_uac_request() will now perform the DNS lookups and directly update the branch dst inside the transaction. - add_uac() must now be called with a special flag during DNS failover (UAC_DNS_FAILOVER). --- modules/tm/t_fwd.c | 176 ++++++++++++++++++++++++++++++++++++---------------- modules/tm/t_fwd.h | 2 + 2 files changed, 125 insertions(+), 53 deletions(-) Diff: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commitdiff;h=44f…

1 0

git:sr_3.0: parse_sip_msg_uri: Log broken URIs only when debugging is enabled.
by Jan Janak 27 Oct '09

27 Oct '09

Module: sip-router Branch: sr_3.0 Commit: 027c5888fcefe56006d6e6cf60335a5c35a45037 URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=027c588… Author: Jan Janak <jan(a)ryngle.com> Committer: Jan Janak <jan(a)ryngle.com> Date: Mon Oct 26 14:15:53 2009 +0100 parse_sip_msg_uri: Log broken URIs only when debugging is enabled. Logging broken Request-URIs with LOG(L_ERR) generates too much traffic in syslog by default. Broken Request-URIs are beyond our control and we should not generate an error message each time we receive and parse one. We log them only when debugging is enabled. --- parser/parse_uri.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/parser/parse_uri.c b/parser/parse_uri.c index 5b29710..741258e 100644 --- a/parser/parse_uri.c +++ b/parser/parse_uri.c @@ -1397,8 +1397,8 @@ int parse_sip_msg_uri(struct sip_msg* msg) tmp_len=msg->first_line.u.request.uri.len; } if (parse_uri(tmp, tmp_len, &msg->parsed_uri)<0){ - LOG(L_ERR, "ERROR: parse_sip_msg_uri: bad uri <%.*s>\n", - tmp_len, tmp); + DBG("ERROR: parse_sip_msg_uri: bad uri <%.*s>\n", + tmp_len, tmp); msg->parsed_uri_ok=0; return -1; }

2 2

SF.net SVN: openser:[5950] branches/1.5/modules/tm/t_hooks.h
by Henning Westerholt 27 Oct '09

27 Oct '09

Revision: 5950 http://openser.svn.sourceforge.net/openser/?rev=5950&view=rev Author: henningw Date: 2009-10-27 15:24:12 +0000 (Tue, 27 Oct 2009) Log Message: ----------- - documentation improvements in tm - add missing explanation for TMCB_REQUEST_BUILT - add some informations about the conditions where a callback returns the complete parsed SIP message - fix a bunch of typos - patch from Timo Reimann, timo dot reimann at 1und1 dot de Modified Paths: -------------- branches/1.5/modules/tm/t_hooks.h This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

1 0

Compiler warnings
by Olle E. Johansson 27 Oct '09

27 Oct '09

Compiling sip-router from clean checkout generates a lot of different warnings. What's the policy about this in the project? SHould we just forget or should we react? In Asterisk we have a configuration flag for ./configure that is called "dev-mode" which forces warnings to generate an error and stop make. This helps motivating developers to fix warnings :-) /O gcc -g -O9 -funroll-loops -Wcast-align -m64 -minline-all-stringops - falign-loops -ftree-vectorize -fno-strict-overflow -mtune=opteron - Wall -DNAME='"ser"' -DVERSION='"2.99.99-pre3"' -DARCH='"x86_64"' - DOS='darwin_' -DOS_QUOTED='"darwin"' -DCOMPILER='"gcc 4.2.1"' - D__CPU_x86_64 -D__OS_darwin -DSER_VER=2099099 -DCFG_DIR='"/usr/local/ etc/ser/"' -DPKG_MALLOC -DSHM_MEM -DSHM_MMAP -DDNS_IP_HACK -DUSE_IPV6 - DUSE_MCAST -DUSE_TCP -DDISABLE_NAGLE -DHAVE_RESOLV_RES -DUSE_DNS_CACHE -DUSE_DNS_FAILOVER -DUSE_DST_BLACKLIST -DUSE_NAPTR -DDBG_QM_MALLOC - DUSE_TLS -DTLS_HOOKS -DFAST_LOCK -DADAPTIVE_WAIT - DADAPTIVE_WAIT_LOOPS=1024 -DCC_GCC_LIKE_ASM -DHAVE_SOCKADDR_SA_LEN - DHAVE_GETHOSTBYNAME2 -DHAVE_UNION_SEMUN -DHAVE_SCHED_YIELD - DHAVE_MSGHDR_MSG_CONTROL -DUSE_ANON_MMAP -DNDEBUG - DHAVE_CONNECT_ECONNRESET_BUG -DHAVE_TIMEGM -DUSE_SIGWAIT -DHAVE_KQUEUE -DHAVE_SELECT -c parser/sdp/sdp.c -o parser/sdp/sdp.o parser/sdp/sdp.c: In function 'parse_sdp': parser/sdp/sdp.c:707: warning: dereferencing type-punned pointer will break strict-aliasing rules parser/sdp/sdp.c:729: warning: dereferencing type-punned pointer will break strict-aliasing rules g

2 1

git:sr_3.0: registrar: Fix handling of cases where contacts > max_contacts
by Jan Janak 27 Oct '09

27 Oct '09

Module: sip-router Branch: sr_3.0 Commit: 1bf989c36a9972389c033b87df0e88c04e56b9c3 URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=1bf989c… Author: Jan Janak <jan(a)ryngle.com> Committer: Jan Janak <jan(a)ryngle.com> Date: Mon Oct 26 16:25:02 2009 +0100 registrar: Fix handling of cases where contacts > max_contacts Registrar should not report an error to syslog if a user exceeds the maximum number of allowed contacts per user. The registrar module sends a reply back with a description of what happened so there is no reason to write this to syslog, max_contacts > 0 is a configuration choice rather than an error. This patch also changes all affected functions in registrar module to indicate that the maximum number of allowed contacts was exceeded by returning a positive number, instead of a negative number to indicate that an error occurred. --- modules_s/registrar/save.c | 38 +++++++++++++++++++++----------------- 1 files changed, 21 insertions(+), 17 deletions(-) diff --git a/modules_s/registrar/save.c b/modules_s/registrar/save.c index 27d134c..fee4548 100644 --- a/modules_s/registrar/save.c +++ b/modules_s/registrar/save.c @@ -257,10 +257,11 @@ static int create_rcv_uri(str** uri, struct sip_msg* m) /* - * Message contained some contacts, but record with same address - * of record was not found so we have to create a new record - * and insert all contacts from the message that have expires - * > 0 + * Message contained some contacts, but record with same address of record was + * not found so we have to create a new record and insert all contacts from + * the message that have expires > 0. The function returns a negative number + * on error, a positive number if the number of contacts would exceed + * max_contacts and 0 on success. */ static inline int insert(struct sip_msg* _m, str* aor, contact_t* _c, udomain_t* _d, str* _u, str *ua, str* aor_filter, int sid) { @@ -289,7 +290,7 @@ static inline int insert(struct sip_msg* _m, str* aor, contact_t* _c, udomain_t* if (max_contacts && (num >= max_contacts)) { rerrno = R_TOO_MANY; ul.delete_urecord(_d, _u); - return -1; + return 1; } num++; @@ -394,7 +395,7 @@ static int test_max_contacts(struct sip_msg* _m, urecord_t* _r, contact_t* _c) _c = get_next_contact(_c); } - DBG("test_max_contacts: %d contacts after commit\n", num); + DBG("test_max_contacts: %d contacts after commit, max_contacts=%d\n", num, max_contacts); if (num > max_contacts) { rerrno = R_TOO_MANY; return 1; @@ -414,6 +415,11 @@ static int test_max_contacts(struct sip_msg* _m, urecord_t* _r, contact_t* _c) * > 0, update the contact * 3) If contact in usrloc exists and expires * == 0, delete contact + * + * The function returns a negative number on error, a positive number if + * max_contacts is set and the number of contacts after the change would + * exceed that maximum number of allowed contacts. On success the function + * returns 0. */ static inline int update(struct sip_msg* _m, urecord_t* _r, str* aor, contact_t* _c, str* _ua, str* aor_filter, int sid) { @@ -434,8 +440,12 @@ static inline int update(struct sip_msg* _m, urecord_t* _r, str* aor, contact_t* if (max_contacts) { ret = test_max_contacts(_m, _r, _c); if (ret != 0) { + /* test_max_contacts returns a negative number on error and a + * positive number if the number of contacts after the update + * exceeds the configured max_contacts. In both cases we return + * here. */ build_contact(_r->contacts, aor_filter); - return -1; + return ret; } } @@ -607,24 +617,18 @@ static inline int contacts(struct sip_msg* _m, contact_t* _c, udomain_t* _d, str } if (res == 0) { /* Contacts found */ - if (update(_m, r, aor, _c, _ua, aor_filter, sid) < 0) { + if ((res = update(_m, r, aor, _c, _ua, aor_filter, sid) < 0)) { LOG(L_ERR, "contacts(): Error while updating record\n"); - build_contact(r->contacts, aor_filter); - ul.release_urecord(r); - ul.unlock_udomain(_d); - return -3; } build_contact(r->contacts, aor_filter); ul.release_urecord(r); } else { - if (insert(_m, aor, _c, _d, _u, _ua, aor_filter, sid) < 0) { + if ((res = insert(_m, aor, _c, _d, _u, _ua, aor_filter, sid) < 0)) { LOG(L_ERR, "contacts(): Error while inserting record\n"); - ul.unlock_udomain(_d); - return -4; } } ul.unlock_udomain(_d); - return 0; + return res; } #define UA_DUMMY_STR "Unknown" @@ -684,7 +688,7 @@ static inline int save_real(struct sip_msg* _m, udomain_t* _t, char* aor_filt, i if (no_contacts(_t, &uid, &aor_filter) < 0) goto error; } } else { - if (contacts(_m, c, _t, &uid, &ua, &aor_filter) < 0) goto error; + if (contacts(_m, c, _t, &uid, &ua, &aor_filter) != 0) goto error; } if (doreply) {

2 2

routing time of sip request
by jh＠tutpro.com 27 Oct '09

27 Oct '09

eu is asking operators to provide all kinds of statistics about their telephone service. one of the attributes is routing time of sip request, i.e., in/out delay of sip requests. in order to be able to do something like that, we could have a pseudo variable that provides ms timestamp value or, perhaps better, a way to zero a ms counter and then store its current value to a pseudo variable in onsend_route. the problems seems to be that according to core cookbook, it is not possible to set pseudo variables in onsend_route. comments? -- juha

6 16

← Newer
1
...
5
6
7
8
9
10
11
...
47
Older →

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

sr-dev October 2009