sr-dev August 2011

sr-dev@lists.kamailio.org

33 participants
291 discussions

git:master: modules_k/presence : Added parameter to allow disabling of sending an initial empty NOTIFY by presence , as this confused some CPEs

by Marius Zbihlei

Module: sip-router Branch: master Commit: aa7b306e24ba1c9b5bdb4e18b88cfc74ad00733b URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=aa7b306… Author: Marius Zbihlei <marius.zbihlei(a)1and1.ro> Committer: Marius Zbihlei <marius.zbihlei(a)1and1.ro> Date: Thu Jun 23 11:42:53 2011 +0300 modules_k/presence : Added parameter to allow disabling of sending an initial empty NOTIFY by presence, as this confused some CPEs Default is enabled, so behavior is left the existing one. Sending of the NOTIFY (carrying the initial/default state) is left now to the application. --- modules_k/presence/README | 63 +++++++++++++++++++---------- modules_k/presence/doc/presence_admin.xml | 21 ++++++++++ modules_k/presence/presence.c | 2 + modules_k/presence/presence.h | 1 + modules_k/presence/subscribe.c | 2 +- 5 files changed, 66 insertions(+), 23 deletions(-) diff --git a/modules_k/presence/README b/modules_k/presence/README index 51ae45e..4384d71 100644 --- a/modules_k/presence/README +++ b/modules_k/presence/README @@ -47,8 +47,9 @@ Juha Heinanen 3.12. db_mode (int) 3.13. subs_htable_size (int) 3.14. pres_htable_size (int) - 3.15. enable_sphere_check (int) - 3.16. timeout_rm_subs (int) + 3.15. send_fast_notify (int) + 3.16. enable_sphere_check (int) + 3.17. timeout_rm_subs (int) 4. Exported Functions @@ -99,13 +100,14 @@ Juha Heinanen 1.12. Set db_mode parameter 1.13. Set subs_htable_size parameter 1.14. Set pres_htable_size parameter - 1.15. Set enable_sphere_check parameter - 1.16. Set timeout_rm_subs parameter - 1.17. handle_publish usage - 1.18. handle_subscribe usage - 1.19. pres_auth_status usage - 1.20. pres_refresh_watchers usage - 1.21. pres_update_watchers usage + 1.15. Set send_fast_notify parameter + 1.16. Set enable_sphere_check parameter + 1.17. Set timeout_rm_subs parameter + 1.18. handle_publish usage + 1.19. handle_subscribe usage + 1.20. pres_auth_status usage + 1.21. pres_refresh_watchers usage + 1.22. pres_update_watchers usage 2.1. presence_api_t structure Chapter 1. Admin Guide @@ -134,8 +136,9 @@ Chapter 1. Admin Guide 3.12. db_mode (int) 3.13. subs_htable_size (int) 3.14. pres_htable_size (int) - 3.15. enable_sphere_check (int) - 3.16. timeout_rm_subs (int) + 3.15. send_fast_notify (int) + 3.16. enable_sphere_check (int) + 3.17. timeout_rm_subs (int) 4. Exported Functions @@ -220,8 +223,9 @@ Chapter 1. Admin Guide 3.12. db_mode (int) 3.13. subs_htable_size (int) 3.14. pres_htable_size (int) - 3.15. enable_sphere_check (int) - 3.16. timeout_rm_subs (int) + 3.15. send_fast_notify (int) + 3.16. enable_sphere_check (int) + 3.17. timeout_rm_subs (int) 3.1. db_url(str) @@ -401,7 +405,22 @@ modparam("presence", "subs_htable_size", 11) modparam("presence", "pres_htable_size", 11) ... -3.15. enable_sphere_check (int) +3.15. send_fast_notify (int) + + This parameter enables or disables the sending of an initial empty + NOTIFY after a SUBSCRIBE/reSUBSCRIBE. This caused problems for MWI + application, because some CPEs (like Samsung) fail to understand an + empty NOTIFY to an message-summary event. This parameter is enabled by + default, thus addering to the standard. + + Default value is “1 ”. + + Example 1.15. Set send_fast_notify parameter +... +modparam("presence", "send_fast_notify", 0) +... + +3.16. enable_sphere_check (int) This parameter is a flag that should be set if permission rules include sphere checking. The sphere information is expected to be present in @@ -411,12 +430,12 @@ modparam("presence", "pres_htable_size", 11) Default value is “0 ”. - Example 1.15. Set enable_sphere_check parameter + Example 1.16. Set enable_sphere_check parameter ... modparam("presence", "enable_sphere_check", 1) ... -3.16. timeout_rm_subs (int) +3.17. timeout_rm_subs (int) This parameter is a flag that should be set if subscriptions should be removed from the active_watchers when a NOTIFY times out. RFC3265 @@ -426,7 +445,7 @@ modparam("presence", "enable_sphere_check", 1) Default value is “1”. - Example 1.16. Set timeout_rm_subs parameter + Example 1.17. Set timeout_rm_subs parameter ... modparam("presence", "timeout_rm_subs", 0) ... @@ -459,7 +478,7 @@ modparam("presence", "timeout_rm_subs", 0) The module sends an appropriate stateless reply in all cases. - Example 1.17. handle_publish usage + Example 1.18. handle_publish usage ... if(is_method("PUBLISH")) { @@ -485,7 +504,7 @@ modparam("presence", "timeout_rm_subs", 0) The module sends an appropriate stateless reply in all cases. - Example 1.18. handle_subscribe usage + Example 1.19. handle_subscribe usage ... if(method=="SUBSCRIBE") handle_subscribe(); @@ -502,7 +521,7 @@ if(method=="SUBSCRIBE") This function can be used from REQUEST_ROUTE. - Example 1.19. pres_auth_status usage + Example 1.20. pres_auth_status usage ... if (method=="MESSAGE") { pres_auth_status("$fu", $ru"); @@ -533,7 +552,7 @@ if (method=="MESSAGE") { This function can be used from ANY_ROUTE. - Example 1.20. pres_refresh_watchers usage + Example 1.21. pres_refresh_watchers usage ... pres_refresh_watchers("sip:test@kamailio.org", "presence", 1); ... @@ -551,7 +570,7 @@ pres_refresh_watchers("sip:test@kamailio.org", "presence", 1); This function can be used from ANY_ROUTE. - Example 1.21. pres_update_watchers usage + Example 1.22. pres_update_watchers usage ... pres_update_watchers("sip:test@kamailio.org", "presence"); ... diff --git a/modules_k/presence/doc/presence_admin.xml b/modules_k/presence/doc/presence_admin.xml index b2a1629..49c8730 100644 --- a/modules_k/presence/doc/presence_admin.xml +++ b/modules_k/presence/doc/presence_admin.xml @@ -359,6 +359,27 @@ modparam("presence", "pres_htable_size", 11) </example> </section> <section> + <title><varname>send_fast_notify</varname> (int)</title> + <para> + This parameter enables or disables the sending of an initial empty NOTIFY after a SUBSCRIBE/reSUBSCRIBE. + This caused problems for MWI application, because some CPEs (like Samsung) fail to understand an empty + NOTIFY to an message-summary event. This parameter is enabled by default, thus addering to the standard. + </para> + <para> + <emphasis>Default value is <quote>1 </quote>. + </emphasis> + </para> + <example> + <title>Set <varname>send_fast_notify</varname> parameter</title> + <programlisting format="linespecific"> +... +modparam("presence", "send_fast_notify", 0) +... + </programlisting> + </example> + </section> + + <section> <title><varname>enable_sphere_check</varname> (int)</title> <para> This parameter is a flag that should be set if permission rules include diff --git a/modules_k/presence/presence.c b/modules_k/presence/presence.c index d211b53..8dd866c 100644 --- a/modules_k/presence/presence.c +++ b/modules_k/presence/presence.c @@ -137,6 +137,7 @@ int dbmode = 0; int fallback2db = 0; int sphere_enable= 0; int timeout_rm_subs = 1; +int send_fast_notify = 1; int phtable_size= 9; phtable_t* pres_htable; @@ -177,6 +178,7 @@ static param_export_t params[]={ { "fallback2db", INT_PARAM, &fallback2db}, { "enable_sphere_check", INT_PARAM, &sphere_enable}, { "timeout_rm_subs", INT_PARAM, &timeout_rm_subs}, + { "send_fast_notify", INT_PARAM, &send_fast_notify}, {0,0,0} }; diff --git a/modules_k/presence/presence.h b/modules_k/presence/presence.h index 88520f3..16e5177 100644 --- a/modules_k/presence/presence.h +++ b/modules_k/presence/presence.h @@ -75,6 +75,7 @@ extern int max_expires; extern int dbmode; extern int sphere_enable; extern int timeout_rm_subs; +extern int send_fast_notify; extern int shtable_size; extern shtable_t subs_htable; diff --git a/modules_k/presence/subscribe.c b/modules_k/presence/subscribe.c index d37518a..3fb1668 100644 --- a/modules_k/presence/subscribe.c +++ b/modules_k/presence/subscribe.c @@ -607,7 +607,7 @@ int update_subscription(struct sip_msg* msg, subs_t* subs, int to_tag_gen, } *sent_reply= 1; - if(notify(subs, NULL, NULL, 0 )< 0) + if(send_fast_notify && (notify(subs, NULL, NULL, 0 )< 0)) { LM_ERR("sending notify request\n"); goto error;

13 years, 4 months

Re: [sr-dev] Patches for siptrace

by Daniel-Constantin Mierla

Hello, let's keep the list on cc so other devs can participate and have idea of what is going on ... On 8/11/11 9:26 AM, Tristan Bruns wrote: > Hello Daniel, > > I tried using the message parser and textops to add the headers to the > message but failed, probably because I did not manage to create a > sip_msg struct that append_hf would update. (It worked when I used an > existing sip_msg struct, [e.g. the one from sip_trace()] but not all > trace-functions had access to a sip_msg [e.g. trace_sl_*]) Indeed, the functions for storing the message don't have the sip_msg struct available. I was referring to parsing of received messages (the mirrored traffic from other kamailios), where would be safer to use the parser -- if there will be another proxy in between mirroring proxy and storage proxy, it can add headers in between the X-Siptrace-..., breaking the read. Anyhow, that is very unlikely -- for the I will go with your patches and update afterwards, when it is more time for it. Cheers, Daniel > > After that, I did not consider to use the message parser to extract > the headers. > > Regards, Tristan > > > Am 09.08.2011 16:26, schrieb Daniel-Constantin Mierla: >> Hi Tristan, >> >> I was looking at the patches and the only thing I want to discuss is >> related to the operations with headers. >> >> You don't use the internal SIP parser, with it you can get the end of >> headers position as well as you can iterate through the list of >> headers and match by name, then take the body and fill local variables. >> >> In the patches (3/5) you use sscanf(...). I would go for using the >> parser, since it is safer in long term. Do you had any special reason >> to go with own kind of parser for SIP and headers? >> >> Cheers, >> Daniel >> >> On 8/5/11 10:17 AM, Tristan Bruns wrote: >>> Hello OpenSER/Kamailio developers, >>> >>> we implemented some additional features in siptrace. >>> The main objective was to have multiple kamailios send duplicates >>> of sent/received packets to one logging server (also running kamailio). >>> We did not just set db_url to the logging server because of >>> performance concerns. >>> >>> The attached patches contain our commits. >>> I hope that you find them useful. >>> >>> Viele Grüße / Best regards, >>> Tristan Bruns (DECOIT GmbH) >>> >>> >>> >>> 0001-modules_k-siptrace-separately-store-to-db-and-or-sen.patch >>> modules_k/siptrace: separately store to db and/or send duplicate >>> >>> 0002-modules_k-siptrace-Add-trace_to_database-configurati.patch >>> modules_k/siptrace: Add trace_to_database configuration parameter >>> >>> Adding configuration parameter to disable writing to the >>> database. >>> We can use this to only duplicate the SIP messages without >>> storing >>> them in our database. >>> >>> 0003-modules_k-siptrace-Add-x-headers-feature.patch >>> modules_k/siptrace: Add "x-headers" feature >>> >>> The "x-headers" feature stores the fromip, toip, method and >>> direction in the message body (using X-* headers). This allows to >>> transmit them using duplicate_uri from one kamailio to an other. >>> >>> 0004-modules_k-siptrace-Add-column-time_us.patch >>> modules_k/siptrace: Add column time_us >>> >>> 0005-modules_k-siptrace-Add-time-to-x-headers.patch >>> modules_k/siptrace: Add time to x-headers >>> >>> >>> >>> >>> >>> _______________________________________________ >>> sr-dev mailing list >>> sr-dev(a)lists.sip-router.org >>> http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-dev >> >> -- >> Daniel-Constantin Mierla --http://www.asipto.com >> Kamailio Advanced Training, Oct 10-13, Berlin:http://asipto.com/u/kat >> http://linkedin.com/in/miconda -- http://twitter.com/miconda > -- Daniel-Constantin Mierla -- http://www.asipto.com Kamailio Advanced Training, Oct 10-13, Berlin: http://asipto.com/u/kat http://linkedin.com/in/miconda -- http://twitter.com/miconda

13 years, 4 months

git:master: modules/db_mysql: Add function declaration to prevent compiler

by Timo Reimann

Module: sip-router Branch: master Commit: eb37a5472fd1c5de7dfbc641fdbe372c0f19fe16 URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=eb37a54… Author: Timo Reimann <timo.reimann(a)1und1.de> Committer: Timo Reimann <timo.reimann(a)1und1.de> Date: Wed Aug 17 11:33:55 2011 +0200 modules/db_mysql: Add function declaration to prevent compiler warning. - Define db_mysql_alloc_buffer(void) in km_dbase.h. --- modules/db_mysql/km_dbase.h | 7 +++++++ 1 files changed, 7 insertions(+), 0 deletions(-) diff --git a/modules/db_mysql/km_dbase.h b/modules/db_mysql/km_dbase.h index c1dec34..a4e7b26 100644 --- a/modules/db_mysql/km_dbase.h +++ b/modules/db_mysql/km_dbase.h @@ -125,4 +125,11 @@ int db_insert_update(const db1_con_t* _h, const db_key_t* _k, const db_val_t* _v int db_mysql_use_table(db1_con_t* _h, const str* _t); +/** + * Allocate a buffer for database module + * No function should be called before this + * \return zero on success, negative value on failure + */ +int db_mysql_alloc_buffer(void); + #endif /* KM_DBASE_H */

13 years, 4 months

git:master: core: added test if SREV_NET_DGRAM_IN callback is set

by Daniel-Constantin Mierla

Module: sip-router Branch: master Commit: 8958a14e8ef86986168f4d555435fe561bb7cb3e URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=8958a14… Author: Daniel-Constantin Mierla <miconda(a)gmail.com> Committer: Daniel-Constantin Mierla <miconda(a)gmail.com> Date: Wed Aug 17 11:29:41 2011 +0200 core: added test if SREV_NET_DGRAM_IN callback is set - patch by Alexandr Dubovikov --- events.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/events.c b/events.c index 048cb45..0b78df8 100644 --- a/events.c +++ b/events.c @@ -188,6 +188,8 @@ int sr_event_enabled(int type) return (_sr_events_list.pkg_set_used!=0)?1:0; case SREV_PKG_SET_REAL_USED: return (_sr_events_list.pkg_set_real_used!=0)?1:0; + case SREV_NET_DGRAM_IN: + return (_sr_events_list.net_dgram_in!=0)?1:0; } return 0; }

13 years, 4 months

git:master: modules_k/dialog: Remove superfluous node check.

by Timo Reimann

Module: sip-router Branch: master Commit: 0e54244acb0064760844e709d9e949e32f827ba9 URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=0e54244… Author: Timo Reimann <timo.reimann(a)1und1.de> Committer: Timo Reimann <timo.reimann(a)1und1.de> Date: Wed Aug 17 11:19:38 2011 +0200 modules_k/dialog: Remove superfluous node check. --- modules_k/dialog/dlg_hash.c | 3 --- 1 files changed, 0 insertions(+), 3 deletions(-) diff --git a/modules_k/dialog/dlg_hash.c b/modules_k/dialog/dlg_hash.c index ed474af..c18cb9a 100644 --- a/modules_k/dialog/dlg_hash.c +++ b/modules_k/dialog/dlg_hash.c @@ -943,9 +943,6 @@ static inline int internal_mi_print_dlg(struct mi_node *rpl, if(node1 == 0) goto error; - if(node1 == 0) - goto error; - if (with_context) { node1 = add_mi_node_child(node, 0, "context", 7, 0, 0); if(node1 == 0)

13 years, 4 months

git:master: modules_k/dialog: Add DLGCB_CONFIRMED_NA callback and make

by Timo Reimann

Module: sip-router Branch: master Commit: e1068cb0ab4377c9debd9aa0cbc398f6dee5c998 URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=e1068cb… Author: Timo Reimann <timo.reimann(a)1und1.de> Committer: Timo Reimann <timo.reimann(a)1und1.de> Date: Wed Aug 17 11:04:40 2011 +0200 modules_k/dialog: Add DLGCB_CONFIRMED_NA callback and make DLGCB_CONFIRM fit accordingly. - Add DLGCB_CONFIRMED_NA callback which is triggered on reception of a 200 OK message (setup-concluding ACK message still pending). - Execute DGLCB_CONFIRMED on reception of a setup-concluding ACK message. - Extend documentation and clarify that DLGCB_REQ_WITHIN does not fire on DLGCB_CONFIRMED. (This is for consistency reasons as BYE requests have their own callback (DLGCB_TERMINATED) as well. Users who wish to track both setup-concluding ACK and other sequential requests must register for both callbacks types now.) - Closes Flyspray issue #143. --- modules_k/dialog/dlg_cb.h | 23 +++++++++-------- modules_k/dialog/dlg_handlers.c | 43 ++++++++++++++++++++------------- modules_k/dialog/doc/dialog_devel.xml | 12 +++++++- 3 files changed, 48 insertions(+), 30 deletions(-) diff --git a/modules_k/dialog/dlg_cb.h b/modules_k/dialog/dlg_cb.h index 9c6fa14..80eae7b 100644 --- a/modules_k/dialog/dlg_cb.h +++ b/modules_k/dialog/dlg_cb.h @@ -64,17 +64,18 @@ typedef str* (*get_dlg_variable_f)( struct dlg_cell* dlg, #define DLGCB_LOADED (1<<0) #define DLGCB_CREATED (1<<1) #define DLGCB_FAILED (1<<2) -#define DLGCB_CONFIRMED (1<<3) -#define DLGCB_REQ_WITHIN (1<<4) -#define DLGCB_TERMINATED (1<<5) -#define DLGCB_EXPIRED (1<<6) -#define DLGCB_EARLY (1<<7) -#define DLGCB_RESPONSE_FWDED (1<<8) -#define DLGCB_RESPONSE_WITHIN (1<<9) -#define DLGCB_MI_CONTEXT (1<<10) -#define DLGCB_RPC_CONTEXT (1<<11) -#define DLGCB_DESTROY (1<<12) -#define DLGCB_SPIRALED (1<<13) +#define DLGCB_CONFIRMED_NA (1<<3) +#define DLGCB_CONFIRMED (1<<4) +#define DLGCB_REQ_WITHIN (1<<5) +#define DLGCB_TERMINATED (1<<6) +#define DLGCB_EXPIRED (1<<7) +#define DLGCB_EARLY (1<<8) +#define DLGCB_RESPONSE_FWDED (1<<9) +#define DLGCB_RESPONSE_WITHIN (1<<10) +#define DLGCB_MI_CONTEXT (1<<11) +#define DLGCB_RPC_CONTEXT (1<<12) +#define DLGCB_DESTROY (1<<13) +#define DLGCB_SPIRALED (1<<14) #define DLGCB_TERMINATED_CONFIRMED (1<<14) struct dlg_callback { diff --git a/modules_k/dialog/dlg_handlers.c b/modules_k/dialog/dlg_handlers.c index 450fd09..c6cd650 100644 --- a/modules_k/dialog/dlg_handlers.c +++ b/modules_k/dialog/dlg_handlers.c @@ -398,7 +398,7 @@ static void dlg_onreply(struct cell* t, int type, struct tmcb_params *param) if (new_state==DLG_STATE_CONFIRMED_NA && old_state!=DLG_STATE_CONFIRMED_NA && old_state!=DLG_STATE_CONFIRMED ) { - LM_DBG("dialog %p confirmed\n",dlg); + LM_DBG("dialog %p confirmed (ACK pending)\n",dlg); if (rpl != FAKED_REPLY) { /* get to tag*/ @@ -445,8 +445,8 @@ static void dlg_onreply(struct cell* t, int type, struct tmcb_params *param) ref_dlg(dlg,1); } - /* dialog confirmed */ - run_dlg_callbacks( DLGCB_CONFIRMED, dlg, req, rpl, DLG_DIR_UPSTREAM, 0); + /* dialog confirmed (ACK pending) */ + run_dlg_callbacks( DLGCB_CONFIRMED_NA, dlg, req, rpl, DLG_DIR_UPSTREAM, 0); if (old_state==DLG_STATE_EARLY) if_update_stat(dlg_enable_stats, early_dlgs, -1); @@ -1168,7 +1168,7 @@ void dlg_onroute(struct sip_msg* req, str *route_params, void *param) if ( (event==DLG_EVENT_REQ || event==DLG_EVENT_REQACK) && new_state==DLG_STATE_CONFIRMED) { - LM_DBG("sequential request successfully processed\n"); + timeout = get_dlg_timeout(req); if (timeout!=default_timeout) { dlg->lifetime = timeout; @@ -1184,20 +1184,29 @@ void dlg_onroute(struct sip_msg* req, str *route_params, void *param) update_dialog_dbinfo(dlg); } - /* within dialog request */ - run_dlg_callbacks( DLGCB_REQ_WITHIN, dlg, req, NULL, dir, 0); - - if ( (event!=DLG_EVENT_REQACK) && - (dlg->cbs.types)&DLGCB_RESPONSE_WITHIN ) { - /* ref the dialog as registered into the transaction callback. - * unref will be done when the callback will be destroyed */ - ref_dlg( dlg , 1); - /* register callback for the replies of this request */ - if ( d_tmb.register_tmcb( req, 0, TMCB_RESPONSE_FWDED, - (dir==DLG_DIR_UPSTREAM)?dlg_seq_down_onreply:dlg_seq_up_onreply, - (void*)dlg, unreference_dialog)<0 ) { - LM_ERR("failed to register TMCB (2)\n"); + if (old_state==DLG_STATE_CONFIRMED_NA) { + LM_DBG("confirming ACK successfully processed\n"); + + /* confirming ACK request */ + run_dlg_callbacks( DLGCB_CONFIRMED, dlg, req, NULL, dir, 0); + } else { + LM_DBG("sequential request successfully processed\n"); + + /* within dialog request */ + run_dlg_callbacks( DLGCB_REQ_WITHIN, dlg, req, NULL, dir, 0); + + if ( (event!=DLG_EVENT_REQACK) && + (dlg->cbs.types)&DLGCB_RESPONSE_WITHIN ) { + /* ref the dialog as registered into the transaction callback. + * unref will be done when the callback will be destroyed */ + ref_dlg( dlg , 1); + /* register callback for the replies of this request */ + if ( d_tmb.register_tmcb( req, 0, TMCB_RESPONSE_FWDED, + (dir==DLG_DIR_UPSTREAM)?dlg_seq_down_onreply:dlg_seq_up_onreply, + (void*)dlg, unreference_dialog)<0 ) { + LM_ERR("failed to register TMCB (2)\n"); unref_dlg( dlg , 1); + } } } } diff --git a/modules_k/dialog/doc/dialog_devel.xml b/modules_k/dialog/doc/dialog_devel.xml index f83a05d..0655894 100644 --- a/modules_k/dialog/doc/dialog_devel.xml +++ b/modules_k/dialog/doc/dialog_devel.xml @@ -51,14 +51,22 @@ </para> </listitem> <listitem> - <para><emphasis>DLGCB_CONFIRMED</emphasis> - called when the + <para><emphasis>DLGCB_CONFIRMED_NA</emphasis> - called when the + dialog is confirmed (2xx replied) but the setup-concluding ACK + message from the caller is yet pending - it's a per dialog type. + </para> + </listitem> + <listitem> + <para><emphasis>DLGCB_CONFIRMED</emphasis> - called when the dialog is confirmed (2xx replied) and the setup-concluding ACK message from the caller has been seen - it's a per dialog type. </para> </listitem> <listitem> <para><emphasis>DLGCB_REQ_WITHIN</emphasis> - called when the - dialog matches a sequential request - it's a per dialog type. + dialog matches a sequential request (excluding setup-concluding + ACK messages which are handled in DLGCB_CONFIRMED) - it's a per + dialog type. </para> </listitem> <listitem>

13 years, 4 months

[tracker] Task closed: 3.x dialog module: dialogs stay in database when in "shutdown only" mode

by sip-router

THIS IS AN AUTOMATED MESSAGE, DO NOT REPLY. The following task is now closed: FS#126 - 3.x dialog module: dialogs stay in database when in "shutdown only" mode User who did this - Timo Reimann (tr) Reason for closing: Fixed Additional comments about closing: Fixed in master (commit 76fc55be0) with proposed solution (clear table after loading dialogs during startup). To be shipped in SR 3.2. More information can be found at the following URL: http://sip-router.org/tracker/index.php?do=details&task_id=126 You are receiving this message because you have requested it from the Flyspray bugtracking system. If you did not expect this message or don't want to receive mails in future, you can change your notification settings at the URL shown above.

13 years, 4 months

git:master: modules_k/dialog: Clear dialog and dialog variable tables after

by Timo Reimann

Module: sip-router Branch: master Commit: 76fc55be0107da80997f9515bb940e837ff899de URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=76fc55b… Author: Timo Reimann <timo.reimann(a)1und1.de> Committer: Timo Reimann <timo.reimann(a)1und1.de> Date: Wed Aug 17 09:58:38 2011 +0200 modules_k/dialog: Clear dialog and dialog variable tables after loading dialogs from database into memory if DB_MODE_SHUTDOWN is used. - Closes Flyspray issue #126 ('3.x dialog module: dialogs stay in database when in "shutdown only" mode'). --- modules_k/dialog/dlg_db_handler.c | 14 ++++++++++++++ 1 files changed, 14 insertions(+), 0 deletions(-) diff --git a/modules_k/dialog/dlg_db_handler.c b/modules_k/dialog/dlg_db_handler.c index 4c283d7..6f55dcf 100644 --- a/modules_k/dialog/dlg_db_handler.c +++ b/modules_k/dialog/dlg_db_handler.c @@ -436,6 +436,13 @@ static int load_dialog_info_from_db(int dlg_hash_size, int fetch_num_rows) }while (nr_rows>0); + if (dlg_db_mode==DB_MODE_SHUTDOWN) { + if (dialog_dbf.delete(dialog_db_handle, 0, 0, 0, 0) < 0) { + LM_ERR("failed to clear dialog table\n"); + goto error; + } + } + end: dialog_dbf.free_result(dialog_db_handle, res); return 0; @@ -546,6 +553,13 @@ static int load_dialog_vars_from_db(int fetch_num_rows) }while (nr_rows>0); + if (dlg_db_mode==DB_MODE_SHUTDOWN) { + if (dialog_dbf.delete(dialog_db_handle, 0, 0, 0, 0) < 0) { + LM_ERR("failed to clear dialog variable table\n"); + goto error; + } + } + end: dialog_dbf.free_result(dialog_db_handle, res); return 0;

13 years, 4 months

[tracker] Comment added: 3.x dialog module: dialogs stay in database when in "shutdown only" mode

by sip-router

THIS IS AN AUTOMATED MESSAGE, DO NOT REPLY. The following task has a new comment added: FS#126 - 3.x dialog module: dialogs stay in database when in "shutdown only" mode User who did this - Henning Westerholt (henningw) ---------- As discussed, I think option 2) with a database clearing directly after the de-serialization on startup is the best option. ---------- More information can be found at the following URL: http://sip-router.org/tracker/index.php?do=details&task_id=126#comment251 You are receiving this message because you have requested it from the Flyspray bugtracking system. If you did not expect this message or don't want to receive mails in future, you can change your notification settings at the URL shown above.

13 years, 4 months

[tracker] Comment added: 3.x dialog module: dialogs stay in database when in "shutdown only" mode

by sip-router

THIS IS AN AUTOMATED MESSAGE, DO NOT REPLY. The following task has a new comment added: FS#126 - 3.x dialog module: dialogs stay in database when in "shutdown only" mode User who did this - Timo Reimann (tr) ---------- Finally, I came to analyze the issue: The reason why it doesn't work is that on dialog module startup, any database-related functionality during run-time is prohibited if the database mode is SHUTDOWN. For that mode, however, there is no code to clear the database after startup; the dialog-processing code in dlg_handlers.c relies on the assumption that dialogs can be deleted from the database during run-time which is not possible in SHUTDOWN mode. This is pretty much what Henning already discovered in the original mailing list thread under http://lists.kamailio.org/pipermail/users/2009-December/025811.html The possible fixes are also there: (1) Make database removal possible during run-time (2) Clear the database either prior to serializing during shutdown or after de-serializing during startup. My personal preference goes towards (2) to have SHUTDOWN mode behave as expected: Access the database only when needed to serialize/de-serialize dialogs without any run-time requirements. ---------- More information can be found at the following URL: http://sip-router.org/tracker/index.php?do=details&task_id=126#comment250 You are receiving this message because you have requested it from the Flyspray bugtracking system. If you did not expect this message or don't want to receive mails in future, you can change your notification settings at the URL shown above.

13 years, 4 months

← Newer
1
...
12
13
14
15
16
17
18
...
30
Older →

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

sr-dev August 2011