Hi again Daniel. Testing other things trying to solve the problem I found that the segmentation fault occurs when Openser is restarted and there are records within location database table. If the table is empty initially, openser starts OK, and users can register without problem. Below is the backtrace of a error when the location table is not empty: #0 0xffffffff7d90e2a0 in preload_udomain (_c=0x10026ad90, _d=0xffffffff5028e9f0) at udomain.c:404 404 if (VAL_NULL(ROW_VALUES(row)) (gdb) bt #0 0xffffffff7d90e2a0 in preload_udomain (_c=0x10026ad90, _d=0xffffffff5028e9f0) at udomain.c:404 #1 0xffffffff7d915844 in child_init (_rank=1) at ul_mod.c:344 #2 0x0000000100082500 in init_mod_child (m=0x1, rank=1, type=0x1000f41b0 "CHILD") at /home/operador/openser-1.2.0-notls//sr_module.c:400 #3 0x0000000100082460 in init_mod_child (m=0x1, rank=1, type=0x1000f41b0 "CHILD") at /home/operador/openser-1.2.0-notls//sr_module.c:394 #4 0x0000000100082460 in init_mod_child (m=0x1, rank=1, type=0x1000f41b0 "CHILD") at /home/operador/openser-1.2.0-notls//sr_module.c:394 #5 0x0000000100082770 in init_child (rank=1) at /home/operador/openser-1.2.0-notls//sr_module.c:394 #6 0x0000000100032414 in main_loop () at /home/operador/openser-1.2.0-notls//main.c:724 #7 0x00000001000345a4 in main (argc=2, argv=0xffffff7eaeeec4ff) at /home/operador/openser-1.2.0-notls//main.c:1399 Thanks. On 3/26/07, *Sergio Gutierrez* &lt;saguti(a)gmail.com <mailto:saguti@gmail.com>> wrote: Hi Daniel. This is the backtrace of the core. Thanks #0 0xffffffff7d90ee9c in preload_udomain (_c=0x10026b7d800, _d=0x0) at /home/operador/openser-1.2.0-notls/modules/usrloc//udomain.c:404 #1 0xffffffff7d9131c4 in child_init (_rank=1344858528) at /home/operador/openser-1.2.0-notls/modules/usrloc//ul_mod.c:344 #2 0x0000000100082500 in init_mod_child (m=0x1, rank=1, type=0x1000f41b0 "CHILD") at /home/operador/openser-1.2.0-notls//sr_module.c:400 #3 0x0000000100082460 in init_mod_child (m=0x1, rank=1, type=0x1000f41b0 "CHILD") at /home/operador/openser-1.2.0-notls//sr_module.c:394 #4 0x0000000100082460 in init_mod_child (m=0x1, rank=1, type=0x1000f41b0 "CHILD") at /home/operador/openser-1.2.0-notls//sr_module.c:394 #5 0x0000000100082770 in init_child (rank=1) at /home/operador/openser-1.2.0-notls//sr_module.c:394 #6 0x0000000100032414 in main_loop () at /home/operador/openser-1.2.0-notls//main.c:724 #7 0x00000001000345a4 in main (argc=2, argv=0xffffff7e7007a000) at /home/operador/openser-1.2.0-notls//main.c:1399 On 3/26/07, *Daniel-Constantin Mierla* < daniel(a)voice-system.ro <mailto:daniel@voice-system.ro>> wrote: Hello, can you send a gdb backtrace :-D -- I cannot read adb backtrace. Cheers, Daniel On 03/26/07 19:42, Sergio Gutierrez wrote: 100269870) 1002699f0) 100269c70) <mailto:saguti@gmail.com> daniel(a)voice-system.ro <mailto:daniel@voice-system.ro> <mailto:daniel@voice-system.ro>>> wrote: parameter -m? SPARC testing radius memory caused by ------------------------------------------------------------------------ <mailto: Users(a)openser.org <mailto:Users@openser.org>> <http://openser.org/cgi-bin/mailman/listinfo/users>

Hi Daniel. I am using mysql version 5.0.37, compiled on 64 bits too. These are the printing of the symbols you asked: (gdb) print row $1 = (db_row_t *) 0x10026bf68 (gdb) print ROW_VALUES(row) No symbol "ROW_VALUES" in current context. (gdb) print VAL_NULL(ROW_VALUES(row) No symbol "VAL_NULL" in current context. (gdb) Thanks. On 3/27/07, *Daniel-Constantin Mierla* &lt;daniel(a)voice-system.ro <mailto:daniel@voice-system.ro>> wrote: Hello Sergio, seems to be some data corrupted from what database returned. What is the database type you use, mysql? Could you print the the content of row and ROW_VALUES(row) (via print command in gdb)? Thanks, Daniel On 03/26/07 23:46, Sergio Gutierrez wrote: empty: <mailto:saguti@gmail.com> 1.2.0-notls/modules/usrloc//udomain.c:404 daniel(a)voice-system.ro <mailto:daniel@voice-system.ro> <mailto:daniel@voice-system.ro>>> wrote: backtrace. 100215a00, 1000f40f0, 1, 100269eb0) 1002694b0, 0) <mailto:saguti@gmail.com> <mailto:saguti@gmail.com <mailto:saguti@gmail.com>>>> wrote: <mailto:daniel@voice-system.ro <mailto:daniel@voice-system.ro>> <mailto:daniel@voice-system.ro> <mailto:daniel@voice-system.ro>>>> wrote: bits on a fault error. error is ------------------------------------------------------------------------ <mailto:Users@openser.org <mailto:Users@openser.org>> <mailto: Users(a)openser.org <mailto:Users@openser.org>>> http://openser.org/cgi-bin/mailman/listinfo/users <http://openser.org/cgi-bin/mailman/listinfo/users>> > > > > > > > > >

Hi Daniel. I apologize for not sending correctly well printed. I am not very proficient with GDB :-) (gdb) print *row $1 = {values = 0x10026bfe8, n = 14} (gdb) print *row->values $2 = {type = DB_INT, nul = 0, val = {int_val = 3950031, double_val = 1.5727205756671771e-307, time_val = 16965253963186176, string_val = 0x3c45cf00000000 <Address 0x3c45cf00000000 out of bounds>, str_val = {s = 0x3c45cf00000000 <Address 0x3c45cf00000000 out of bounds>, len = 1}, blob_val = {s = 0x3c45cf00000000 <Address 0x3c45cf00000000 out of bounds>, len = 1}, bitmap_val = 3950031}} Thanks. Sergio Gutiérrez EPM Telecomunicaciones S.A. E.S.P Medellin, Colombia Tel: 574-3950022 On 3/29/07, *Daniel-Constantin Mierla* &lt;daniel(a)voice-system.ro <mailto:daniel@voice-system.ro>> wrote: Hello, do not help too much, please print the content of the variables: print *row print *row->values Thanks, Daniel

Hi Daniel. There it goes: mysql> desc location; +---------------+------------------+------+-----+---------------------+----------------+ | Field | Type | Null | Key | Default | Extra | +---------------+------------------+------+-----+---------------------+----------------+ | id | int(10) unsigned | NO | PRI | NULL | auto_increment | | username | varchar(64) | NO | MUL | | | | domain | varchar(128) | NO | | | | | contact | varchar(255) | NO | | | | | received | varchar(255) | YES | | NULL | | | path | varchar(255) | YES | | NULL | | | expires | datetime | NO | | 2020-05-28 21:32:15 | | | q | float(10,2) | NO | | 1.00 | | | callid | varchar(255) | NO | | Default-Call-ID | | | cseq | int(11) | NO | | 13 | | | last_modified | datetime | NO | | 1900-01-01 00:00:00 | | | flags | int(11) | NO | | 0 | | | cflags | int(11) | NO | | 0 | | | user_agent | varchar(255) | NO | | | | | socket | varchar(128) | YES | | NULL | | | methods | int(11) | YES | | NULL | | +---------------+------------------+------+-----+---------------------+----------------+ | location | CREATE TABLE `location` ( `id` int(10) unsigned NOT NULL auto_increment, `username` varchar(64) NOT NULL default '', `domain` varchar(128) NOT NULL default '', `contact` varchar(255) NOT NULL default '', `received` varchar(255) default NULL, `path` varchar(255) default NULL, `expires` datetime NOT NULL default '2020-05-28 21:32:15', `q` float(10,2) NOT NULL default '1.00', `callid` varchar(255) NOT NULL default 'Default-Call-ID', `cseq` int(11) NOT NULL default '13', `last_modified` datetime NOT NULL default '1900-01-01 00:00:00', `flags` int(11) NOT NULL default '0', `cflags` int(11) NOT NULL default '0', `user_agent` varchar(255) NOT NULL default '', `socket` varchar(128) default NULL, `methods` int(11) default NULL, PRIMARY KEY (`id`), KEY `udc_loc` (`username`,`domain`,`contact`) ) ENGINE=MyISAM AUTO_INCREMENT=14 DEFAULT CHARSET=latin1 | mysql> select * from location; +----+----------+--------+---------------------------------+----------+------+---------------------+-------+-------------------------------------------+------+---------------------+-------+--------+--------------------------------------------+--------------------------+---------+ | id | username | domain | contact | received | path | expires | q | callid | cseq | last_modified | flags | cflags | user_agent | socket | methods | +----+----------+--------+---------------------------------+----------+------+---------------------+-------+-------------------------------------------+------+---------------------+-------+--------+--------------------------------------------+--------------------------+---------+ | 13 | 3950031 | | sip:3950031@192.168.198.10:5060 | NULL | NULL | 2007-03-27 00:46:08 | -1.00 | 5e897d38-ce5af712-6f3a7b51(a)192.168.198.10 <mailto:5e897d38-ce5af712-6f3a7b51@192.168.198.10> | 75 | 2007-03-27 00:45:08 | 0 | 0 | PolycomSoundPointIP-SPIP_300-UA/1.6.5.0043 | udp:192.168.199.200:5060 <http://192.168.199.200:5060> | 8063 | +----+----------+--------+---------------------------------+----------+------+---------------------+-------+-------------------------------------------+------+---------------------+-------+--------+--------------------------------------------+--------------------------+---------+ The last information is the record I have stored. Kind regards.

Hi Daniel. First, I have to apologize because I made you lost some time about this issue. Following a recommendation of Jeffrey because something we detected trying to diagnose the problem of snmpstat module, I recompiled and instaled from scratch the whole componentes of my instalation: MySQL, Libradiusclient and snmp; later Openser. At my previous test, I recompiled individual modules, or I used "non clean" source directories, and, in the same way that happened with snmpstats, I think that the error which was causing the segmentation fault could be produced by an inconsistent linking on the compilation process. At this time, I can confirm that Openser is working perfectly compiled on Solaris SPARC hardware, on 64 bits, and using the recommended optimizations for the package GCC for Sparc Systems. There were just a couple of things I had to have present at compilation time: 1. I included within my Solaris installation the Freeware packages, which include libxml. So, for the compilation, I had to include in CFLAGS the directory of includes (/usr/sfw/include) and in LDFLAGS the libraries directory (/usr/sfw/lib/sparcv). Would it be possible to include within makefiles this path as a possible header/libraries search path, on Solaris Platforms? 2. I commented out the -Wcast-align option, avoiding a bunch of warnings about cast alignment. If you consider it worth, I can publish some lines about these optimizations. For the results of my performance tests, I found an improvement of Openser behaviour when compiled using this optimized version of GCC. Again, I apologize for my mistake. Thank you Daniel, and thanks to Jeffrey by his guidance. Kind regards. Sergio Gutiérrez. On 3/30/07, *Daniel-Constantin Mierla* &lt;daniel(a)voice-system.ro <mailto:daniel@voice-system.ro>> wrote: Hello Sergio, very strange that the column type gets DB_INT. Could you apply attached patch (just prints the type of mysql field), compile and install openser, run in debug=7 and send to me the log. I would need the version of mysql client library you have installed to match the value printed in log. Thanks, Daniel On 03/30/07 19:59, Sergio Gutierrez wrote: +---------------+------------------+------+-----+---------------------+----------------+ +---------------+------------------+------+-----+---------------------+----------------+ 21:32:15 00:00:00 +---------------+------------------+------+-----+---------------------+----------------+ +----+----------+--------+---------------------------------+----------+------+---------------------+-------+-------------------------------------------+------+---------------------+-------+--------+--------------------------------------------+--------------------------+---------+ received +----+----------+--------+---------------------------------+----------+------+---------------------+-------+-------------------------------------------+------+---------------------+-------+--------+--------------------------------------------+--------------------------+---------+ <mailto:5e897d38-ce5af712-6f3a7b51@192.168.198.10> <mailto:5e897d38-ce5af712-6f3a7b51@192.168.198.10>> | 75 | 2007-03-27 PolycomSoundPointIP-SPIP_300-UA/1.6.5.0043 <http://192.168.199.200:5060> | 8063 | +----+----------+--------+---------------------------------+----------+------+---------------------+-------+-------------------------------------------+------+---------------------+-------+--------+--------------------------------------------+--------------------------+---------+ > The last information is the record I have stored. > Kind regards.