Re: [Sofia-sip-devel] test_sip_events.c:338: test_nua test_events() FAILED
Hi, > -Original Message- > From: ext Yungwei Chen [mailto:yung...@resolvity.com] > Sent: Friday, April 15, 2011 7:58 PM > To: sofia-sip-devel@lists.sourceforge.net > Subject: [Sofia-sip-devel] test_sip_events.c:338: test_nua > test_events() FAILED > > I am trying to compile sofia-sip-1.12.10 on CentOS 5.5, and "make > check" shows the following. How can I fix it? Thanks. > > TEST NUA-12.5: un-SUBSCRIBE > test_sip_events.c:338: test_nua test_events() FAILED: tl_find(e->data- > >e_tags, nutag_substate)->t_value != nua_substate_terminated or 2 != 3 I have a similar problem popping up in a different test (test_simple) off a slightly modified master tip. I suspect there is a race condition there somewhere. There is certainly a problem in s2check based tests, where the "time warp" hook is not used with regard to thread scheduling, but this is a different part of the test suite. Best regards, Mikhail -- Benefiting from Server Virtualization: Beyond Initial Workload Consolidation -- Increasing the use of server virtualization is a top priority.Virtualization can reduce costs, simplify management, and improve application availability and disaster protection. Learn more about boosting the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev ___ Sofia-sip-devel mailing list Sofia-sip-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/sofia-sip-devel
Re: [Sofia-sip-devel] NUTAG_DETECT_NETWORK_UPDATES still unimplemented under linux?
Al 17/04/11 19:35, En/na Luca Olivetti ha escrit:
> A2 = MD5(REGISTER:sip:ekiga.net)
> auth_response: db79b8a5046c76654b742311b7d4a33d =
> MD5(53060310bbad90036d54e1954983ae4e:4dab21dbe3306f1513e34e9e24eedc00a7827bd9394f:14bd7ed1967527d3770ab1f7c12901c5)
> (qop=NONE)
> svd: nta.c: 7785: outgoing_create: Assertion
> `tport_name_is_resolved(orq->orq_tpn)' failed.
>
It seems the same as this one:
http://www.mail-archive.com/sofia-sip-devel@lists.sourceforge.net/msg03243.html
outgoing_create is using a "bogus" override_tport, here, around line 7770 of
nua.c,
causing the crash
/* select the tport to use for the outgoing message */
if (override_tport) {
/* note: no ref taken to the tport as its only used once here */
if (tport_is_secondary(override_tport)) {
tpn = tport_name(override_tport); <<
orq->orq_user_tport = 1;
}
}
if (tpn) {
/* CANCEL or ACK to [3456]XX */
invalid = tport_name_dup(home, orq->orq_tpn, tpn);
#if HAVE_SOFIA_SRESOLV
/* We send ACK or CANCEL only if original request was really sent */
assert(tport_name_is_resolved(orq->orq_tpn));
#endif
resolved = tport_name_is_resolved(orq->orq_tpn);
orq->orq_url = url_hdup(home, sip->sip_request->rq_url);
In the line marked with << tpn is assigned an empty (not NULL) string,
hence the
failed assert.
Where does it come from, I don't know.
Checking that tpn is not empty just makes the stack crash later on:
nta.c: 2541: outgoing_insert_via: Assertion `via' failed.
Bye
--
Luca
--
Benefiting from Server Virtualization: Beyond Initial Workload
Consolidation -- Increasing the use of server virtualization is a top
priority.Virtualization can reduce costs, simplify management, and improve
application availability and disaster protection. Learn more about boosting
the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev
___
Sofia-sip-devel mailing list
Sofia-sip-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sofia-sip-devel
Re: [Sofia-sip-devel] NUTAG_DETECT_NETWORK_UPDATES still unimplemented under linux?
Hi Luca, 2011/4/18 Luca Olivetti : >> A2 = MD5(REGISTER:sip:ekiga.net) >> auth_response: db79b8a5046c76654b742311b7d4a33d = >> MD5(53060310bbad90036d54e1954983ae4e:4dab21dbe3306f1513e34e9e24eedc00a7827bd9394f:14bd7ed1967527d3770ab1f7c12901c5) >> (qop=NONE) >> svd: nta.c: 7785: outgoing_create: Assertion >> `tport_name_is_resolved(orq->orq_tpn)' failed. > > It seems the same as this one: > > http://www.mail-archive.com/sofia-sip-devel@lists.sourceforge.net/msg03243.html > > outgoing_create is using a "bogus" override_tport, here, around line 7770 of > nua.c, > causing the crash The network change is probably suffering from bit-rot. The transport is cached in nua/outbound.c, I suppose it should drop its reference to tport_t, too, when network change is detected. See: https://gitorious.org/~ppessi/sofia-sip/pessi-sofia-sip/commits/network-change -- Pekka.Pessi mail at nokia.com -- Benefiting from Server Virtualization: Beyond Initial Workload Consolidation -- Increasing the use of server virtualization is a top priority.Virtualization can reduce costs, simplify management, and improve application availability and disaster protection. Learn more about boosting the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev ___ Sofia-sip-devel mailing list Sofia-sip-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/sofia-sip-devel
Re: [Sofia-sip-devel] NUTAG_DETECT_NETWORK_UPDATES still unimplemented under linux?
Al 18/04/11 23:52, En/na Pekka Pessi ha escrit:
> Hi Luca,
>
> 2011/4/18 Luca Olivetti :
>>> A2 = MD5(REGISTER:sip:ekiga.net)
>>> auth_response: db79b8a5046c76654b742311b7d4a33d =
>>> MD5(53060310bbad90036d54e1954983ae4e:4dab21dbe3306f1513e34e9e24eedc00a7827bd9394f:14bd7ed1967527d3770ab1f7c12901c5)
>>> (qop=NONE)
>>> svd: nta.c: 7785: outgoing_create: Assertion
>>> `tport_name_is_resolved(orq->orq_tpn)' failed.
>>
>> It seems the same as this one:
>>
>> http://www.mail-archive.com/sofia-sip-devel@lists.sourceforge.net/msg03243.html
>>
>> outgoing_create is using a "bogus" override_tport, here, around line 7770 of
>> nua.c,
>> causing the crash
>
> The network change is probably suffering from bit-rot.
>
> The transport is cached in nua/outbound.c, I suppose it should drop
> its reference to tport_t, too, when network change is detected. See:
> https://gitorious.org/~ppessi/sofia-sip/pessi-sofia-sip/commits/network-change
Thank you, it doesn't crash anymore.
What about the other issues?
I'm referring to the fact that su_root_add_network_changed is called 3
times, 3 threads are activated and for each address change 3 (or 6, if
monitoring for RTM_DELADDR) network changes will be triggered.
Now it doesn't crash, but it isn't pretty.
Probably for the RTM_NEWADDR/RTM_DELADDR it's my fault (I should somehow
coalesce them into one event).
I could also keep the workaround to start just one thread, i.e.
static void *su_start_nw_os_thread(void *ptr)
{
static int already = 0;
if (already)
return NULL;
already=1;
su_network_changed_t *snc = (su_network_changed_t *) ptr;
but it's ugly and the OS X code isn't doing it.
Bye
--
Luca
--
Benefiting from Server Virtualization: Beyond Initial Workload
Consolidation -- Increasing the use of server virtualization is a top
priority.Virtualization can reduce costs, simplify management, and improve
application availability and disaster protection. Learn more about boosting
the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev
___
Sofia-sip-devel mailing list
Sofia-sip-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sofia-sip-devel
