[bug] PJSIP 2.5.5 with PJ_HAS_IPV6=1: pjsip_endpt_create_resolver failed on system without IPv6

AG
Alexei Gradinari
Tue, Aug 9, 2016 9:20 PM

Hello,

The function init_sock at pjlib-util/src/pjlib-util/resolver.c
should check if the IPv6 is available instead of return an error.

Regards,
Alexei

Hello, The function init_sock at pjlib-util/src/pjlib-util/resolver.c should check if the IPv6 is available instead of return an error. Regards, Alexei
NI
Nanang Izzuddin
Tue, Aug 16, 2016 6:45 AM

Hi Alexei,

Thank you for the report. The bug sounds valid, unfortunately I could not
reproduce the issue on Windows 10 and Linux, have tried to manually disable
IPv6 but still init_sock() did not fail (perhaps IPv6 loopback interface is
still up). So I am afraid we'd need some more help to test the attached
patch, thanks again in advance.

BR,
nanang

On Wed, Aug 10, 2016 at 4:20 AM, Alexei Gradinari alex2grad@gmail.com
wrote:

Hello,

The function init_sock at pjlib-util/src/pjlib-util/resolver.c
should check if the IPv6 is available instead of return an error.

Regards,
Alexei


Visit our blog: http://blog.pjsip.org

pjsip mailing list
pjsip@lists.pjsip.org
http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org

Hi Alexei, Thank you for the report. The bug sounds valid, unfortunately I could not reproduce the issue on Windows 10 and Linux, have tried to manually disable IPv6 but still init_sock() did not fail (perhaps IPv6 loopback interface is still up). So I am afraid we'd need some more help to test the attached patch, thanks again in advance. BR, nanang On Wed, Aug 10, 2016 at 4:20 AM, Alexei Gradinari <alex2grad@gmail.com> wrote: > Hello, > > The function init_sock at pjlib-util/src/pjlib-util/resolver.c > should check if the IPv6 is available instead of return an error. > > Regards, > Alexei > > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >
AG
Alexei Gradinari
Tue, Aug 16, 2016 2:15 PM

Re: [pjsip] [bug] PJSIP 2.5.5 with PJ_HAS_IPV6=1: pjsip_endpt_create_resolver failed on system without IPv6  Hello Nanang,

I tested on Linux when kernel compiled without IPv6

CONFIG_IPV6 is not set

Regards,
Alexei

Tuesday, August 16, 2016, 2:45:58 AM, you wrote:

Hi Alexei,

Thank you for the report. The bug sounds valid, unfortunately I could not reproduce the issue on Windows 10 and Linux, have tried to manually disable IPv6 but still init_sock() did not fail (perhaps IPv6 loopback interface is still up). So I am afraid we'd need some more help to test the attached patch, thanks again in advance.

BR,
nanang

On Wed, Aug 10, 2016 at 4:20 AM, Alexei Gradinari <alex2grad@gmail.com> wrote:
Hello,

The function init_sock at pjlib-util/src/pjlib-util/resolver.c
should check if the IPv6 is available instead of return an error.

Regards,
Alexei

_______________________________________________
Visit our blog: http://blog.pjsip.org

pjsip mailing list
pjsip@lists.pjsip.org
http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org

GJ
George Joseph
Tue, Aug 16, 2016 3:50 PM

I can test.  I ran into this issue yesterday.  If it's the same error, it
happened for me when the name server returned a normal response for the A
record but a SERVER FAIL response for the AAAA record.  In this case we got
"dns not working" failures from pjproject.

On Tue, Aug 16, 2016 at 12:45 AM, Nanang Izzuddin nanang@pjsip.org wrote:

Hi Alexei,

Thank you for the report. The bug sounds valid, unfortunately I could not
reproduce the issue on Windows 10 and Linux, have tried to manually disable
IPv6 but still init_sock() did not fail (perhaps IPv6 loopback interface is
still up). So I am afraid we'd need some more help to test the attached
patch, thanks again in advance.

BR,
nanang

On Wed, Aug 10, 2016 at 4:20 AM, Alexei Gradinari alex2grad@gmail.com
wrote:

Hello,

The function init_sock at pjlib-util/src/pjlib-util/resolver.c
should check if the IPv6 is available instead of return an error.

Regards,
Alexei


Visit our blog: http://blog.pjsip.org

pjsip mailing list
pjsip@lists.pjsip.org
http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org

--
George Joseph
Digium, Inc. | Software Developer
445 Jan Davis Drive NW - Huntsville, AL 35806 - US
Check us out at: www.digium.com & www.asterisk.org

I can test. I ran into this issue yesterday. If it's the same error, it happened for me when the name server returned a normal response for the A record but a SERVER FAIL response for the AAAA record. In this case we got "dns not working" failures from pjproject. On Tue, Aug 16, 2016 at 12:45 AM, Nanang Izzuddin <nanang@pjsip.org> wrote: > Hi Alexei, > > Thank you for the report. The bug sounds valid, unfortunately I could not > reproduce the issue on Windows 10 and Linux, have tried to manually disable > IPv6 but still init_sock() did not fail (perhaps IPv6 loopback interface is > still up). So I am afraid we'd need some more help to test the attached > patch, thanks again in advance. > > BR, > nanang > > > On Wed, Aug 10, 2016 at 4:20 AM, Alexei Gradinari <alex2grad@gmail.com> > wrote: > >> Hello, >> >> The function init_sock at pjlib-util/src/pjlib-util/resolver.c >> should check if the IPv6 is available instead of return an error. >> >> Regards, >> Alexei >> >> >> _______________________________________________ >> Visit our blog: http://blog.pjsip.org >> >> pjsip mailing list >> pjsip@lists.pjsip.org >> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >> > > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > > -- George Joseph Digium, Inc. | Software Developer 445 Jan Davis Drive NW - Huntsville, AL 35806 - US Check us out at: www.digium.com & www.asterisk.org
NI
Nanang Izzuddin
Wed, Aug 17, 2016 4:33 AM

Hi Alexei,

Finally have managed to reproduce the issue and test the patch here, the
patch seems to work, so just checked in it (with slight modification) to
SVN trunk for ticket https://trac.pjsip.org/repos/ticket/1953.

BR,
nanang

On Tue, Aug 16, 2016 at 9:15 PM, Alexei Gradinari alex2grad@gmail.com
wrote:

Hello Nanang,

I tested on Linux when kernel compiled without IPv6

CONFIG_IPV6 is not set

Regards,
Alexei

Tuesday, August 16, 2016, 2:45:58 AM, you wrote:

Hi Alexei,

Thank you for the report. The bug sounds valid, unfortunately I could not
reproduce the issue on Windows 10 and Linux, have tried to manually disable
IPv6 but still init_sock() did not fail (perhaps IPv6 loopback interface is
still up). So I am afraid we'd need some more help to test the attached
patch, thanks again in advance.

BR,
nanang

On Wed, Aug 10, 2016 at 4:20 AM, Alexei Gradinari alex2grad@gmail.com
wrote:
Hello,

The function init_sock at pjlib-util/src/pjlib-util/resolver.c
should check if the IPv6 is available instead of return an error.

Regards,
Alexei


Visit our blog: http://blog.pjsip.org

pjsip mailing list
pjsip@lists.pjsip.org
http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org

Hi Alexei, Finally have managed to reproduce the issue and test the patch here, the patch seems to work, so just checked in it (with slight modification) to SVN trunk for ticket https://trac.pjsip.org/repos/ticket/1953. BR, nanang On Tue, Aug 16, 2016 at 9:15 PM, Alexei Gradinari <alex2grad@gmail.com> wrote: > Hello Nanang, > > I tested on Linux when kernel compiled without IPv6 > # CONFIG_IPV6 is not set > > Regards, > Alexei > > > Tuesday, August 16, 2016, 2:45:58 AM, you wrote: > > Hi Alexei, > > Thank you for the report. The bug sounds valid, unfortunately I could not > reproduce the issue on Windows 10 and Linux, have tried to manually disable > IPv6 but still init_sock() did not fail (perhaps IPv6 loopback interface is > still up). So I am afraid we'd need some more help to test the attached > patch, thanks again in advance. > > BR, > nanang > > > On Wed, Aug 10, 2016 at 4:20 AM, Alexei Gradinari <alex2grad@gmail.com> > wrote: > Hello, > > The function init_sock at pjlib-util/src/pjlib-util/resolver.c > should check if the IPv6 is available instead of return an error. > > Regards, > Alexei > > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >
NI
Nanang Izzuddin
Wed, Aug 17, 2016 4:41 AM

Hi George,

Perhaps your case is not related to this issue, please also check ticket
#1953 for more detail.

BR,
nanang

On Tue, Aug 16, 2016 at 10:50 PM, George Joseph gjoseph@digium.com wrote:

I can test.  I ran into this issue yesterday.  If it's the same error, it
happened for me when the name server returned a normal response for the A
record but a SERVER FAIL response for the AAAA record.  In this case we got
"dns not working" failures from pjproject.

On Tue, Aug 16, 2016 at 12:45 AM, Nanang Izzuddin nanang@pjsip.org
wrote:

Hi Alexei,

Thank you for the report. The bug sounds valid, unfortunately I could not
reproduce the issue on Windows 10 and Linux, have tried to manually disable
IPv6 but still init_sock() did not fail (perhaps IPv6 loopback interface is
still up). So I am afraid we'd need some more help to test the attached
patch, thanks again in advance.

BR,
nanang

On Wed, Aug 10, 2016 at 4:20 AM, Alexei Gradinari alex2grad@gmail.com
wrote:

Hello,

The function init_sock at pjlib-util/src/pjlib-util/resolver.c
should check if the IPv6 is available instead of return an error.

Regards,
Alexei


Visit our blog: http://blog.pjsip.org

pjsip mailing list
pjsip@lists.pjsip.org
http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org

--
George Joseph
Digium, Inc. | Software Developer
445 Jan Davis Drive NW - Huntsville, AL 35806 - US
Check us out at: www.digium.com & www.asterisk.org


Visit our blog: http://blog.pjsip.org

pjsip mailing list
pjsip@lists.pjsip.org
http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org

Hi George, Perhaps your case is not related to this issue, please also check ticket #1953 for more detail. BR, nanang On Tue, Aug 16, 2016 at 10:50 PM, George Joseph <gjoseph@digium.com> wrote: > I can test. I ran into this issue yesterday. If it's the same error, it > happened for me when the name server returned a normal response for the A > record but a SERVER FAIL response for the AAAA record. In this case we got > "dns not working" failures from pjproject. > > > On Tue, Aug 16, 2016 at 12:45 AM, Nanang Izzuddin <nanang@pjsip.org> > wrote: > >> Hi Alexei, >> >> Thank you for the report. The bug sounds valid, unfortunately I could not >> reproduce the issue on Windows 10 and Linux, have tried to manually disable >> IPv6 but still init_sock() did not fail (perhaps IPv6 loopback interface is >> still up). So I am afraid we'd need some more help to test the attached >> patch, thanks again in advance. >> >> BR, >> nanang >> >> >> On Wed, Aug 10, 2016 at 4:20 AM, Alexei Gradinari <alex2grad@gmail.com> >> wrote: >> >>> Hello, >>> >>> The function init_sock at pjlib-util/src/pjlib-util/resolver.c >>> should check if the IPv6 is available instead of return an error. >>> >>> Regards, >>> Alexei >>> >>> >>> _______________________________________________ >>> Visit our blog: http://blog.pjsip.org >>> >>> pjsip mailing list >>> pjsip@lists.pjsip.org >>> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >>> >> >> >> _______________________________________________ >> Visit our blog: http://blog.pjsip.org >> >> pjsip mailing list >> pjsip@lists.pjsip.org >> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >> >> > > > -- > George Joseph > Digium, Inc. | Software Developer > 445 Jan Davis Drive NW - Huntsville, AL 35806 - US > Check us out at: www.digium.com & www.asterisk.org > > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > >
GJ
George Joseph
Wed, Aug 17, 2016 12:46 PM

On Tue, Aug 16, 2016 at 10:41 PM, Nanang Izzuddin nanang@pjsip.org wrote:

Hi George,

Perhaps your case is not related to this issue, please also check ticket
#1953 for more detail.

Ah, you're right.  I hadn't actually looked. :)
Separate issue then and I'll look deeper.

BR,
nanang

On Tue, Aug 16, 2016 at 10:50 PM, George Joseph gjoseph@digium.com
wrote:

I can test.  I ran into this issue yesterday.  If it's the same error, it
happened for me when the name server returned a normal response for the A
record but a SERVER FAIL response for the AAAA record.  In this case we got
"dns not working" failures from pjproject.

On Tue, Aug 16, 2016 at 12:45 AM, Nanang Izzuddin nanang@pjsip.org
wrote:

Hi Alexei,

Thank you for the report. The bug sounds valid, unfortunately I could
not reproduce the issue on Windows 10 and Linux, have tried to manually
disable IPv6 but still init_sock() did not fail (perhaps IPv6 loopback
interface is still up). So I am afraid we'd need some more help to test the
attached patch, thanks again in advance.

BR,
nanang

On Wed, Aug 10, 2016 at 4:20 AM, Alexei Gradinari alex2grad@gmail.com
wrote:

Hello,

The function init_sock at pjlib-util/src/pjlib-util/resolver.c
should check if the IPv6 is available instead of return an error.

Regards,
Alexei


Visit our blog: http://blog.pjsip.org

pjsip mailing list
pjsip@lists.pjsip.org
http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org

--
George Joseph
Digium, Inc. | Software Developer
445 Jan Davis Drive NW - Huntsville, AL 35806 - US
Check us out at: www.digium.com & www.asterisk.org


Visit our blog: http://blog.pjsip.org

pjsip mailing list
pjsip@lists.pjsip.org
http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org

--
George Joseph
Digium, Inc. | Software Developer
445 Jan Davis Drive NW - Huntsville, AL 35806 - US
Check us out at: www.digium.com & www.asterisk.org

On Tue, Aug 16, 2016 at 10:41 PM, Nanang Izzuddin <nanang@pjsip.org> wrote: > Hi George, > > Perhaps your case is not related to this issue, please also check ticket > #1953 for more detail. > Ah, you're right. I hadn't actually looked. :) Separate issue then and I'll look deeper. > > BR, > nanang > > > On Tue, Aug 16, 2016 at 10:50 PM, George Joseph <gjoseph@digium.com> > wrote: > >> I can test. I ran into this issue yesterday. If it's the same error, it >> happened for me when the name server returned a normal response for the A >> record but a SERVER FAIL response for the AAAA record. In this case we got >> "dns not working" failures from pjproject. >> >> >> On Tue, Aug 16, 2016 at 12:45 AM, Nanang Izzuddin <nanang@pjsip.org> >> wrote: >> >>> Hi Alexei, >>> >>> Thank you for the report. The bug sounds valid, unfortunately I could >>> not reproduce the issue on Windows 10 and Linux, have tried to manually >>> disable IPv6 but still init_sock() did not fail (perhaps IPv6 loopback >>> interface is still up). So I am afraid we'd need some more help to test the >>> attached patch, thanks again in advance. >>> >>> BR, >>> nanang >>> >>> >>> On Wed, Aug 10, 2016 at 4:20 AM, Alexei Gradinari <alex2grad@gmail.com> >>> wrote: >>> >>>> Hello, >>>> >>>> The function init_sock at pjlib-util/src/pjlib-util/resolver.c >>>> should check if the IPv6 is available instead of return an error. >>>> >>>> Regards, >>>> Alexei >>>> >>>> >>>> _______________________________________________ >>>> Visit our blog: http://blog.pjsip.org >>>> >>>> pjsip mailing list >>>> pjsip@lists.pjsip.org >>>> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >>>> >>> >>> >>> _______________________________________________ >>> Visit our blog: http://blog.pjsip.org >>> >>> pjsip mailing list >>> pjsip@lists.pjsip.org >>> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >>> >>> >> >> >> -- >> George Joseph >> Digium, Inc. | Software Developer >> 445 Jan Davis Drive NW - Huntsville, AL 35806 - US >> Check us out at: www.digium.com & www.asterisk.org >> >> >> _______________________________________________ >> Visit our blog: http://blog.pjsip.org >> >> pjsip mailing list >> pjsip@lists.pjsip.org >> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >> >> > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > > -- George Joseph Digium, Inc. | Software Developer 445 Jan Davis Drive NW - Huntsville, AL 35806 - US Check us out at: www.digium.com & www.asterisk.org
AG
Alexei Gradinari
Wed, Aug 17, 2016 3:17 PM

Re: [pjsip] [bug] PJSIP 2.5.5 with PJ_HAS_IPV6=1: pjsip_endpt_create_resolver failed on system without IPv6  Hello Nanang,

Can confirm this patched fixed the issue.
Could you, please, add logging of the PJ_HAS_IPV6 in pj_dump_config.

Regards,
Alexei

Wednesday, August 17, 2016, 12:33:55 AM, you wrote:

Hi Alexei,

Finally have managed to reproduce the issue and test the patch here, the patch seems to work, so just checked in it (with slight modification) to SVN trunk for ticket https://trac.pjsip.org/repos/ticket/1953.

BR,
nanang

On Tue, Aug 16, 2016 at 9:15 PM, Alexei Gradinari <alex2grad@gmail.com> wrote:
Hello Nanang,

I tested on Linux when kernel compiled without IPv6

CONFIG_IPV6 is not set

Regards,
Alexei

Tuesday, August 16, 2016, 2:45:58 AM, you wrote:

Hi Alexei,

Thank you for the report. The bug sounds valid, unfortunately I could not reproduce the issue on Windows 10 and Linux, have tried to manually disable IPv6 but still init_sock() did not fail (perhaps IPv6 loopback interface is still up). So I am afraid we'd need some more help to test the attached patch, thanks again in advance.

BR,
nanang

On Wed, Aug 10, 2016 at 4:20 AM, Alexei Gradinari <alex2grad@gmail.com> wrote:
Hello,

The function init_sock at pjlib-util/src/pjlib-util/resolver.c
should check if the IPv6 is available instead of return an error.

Regards,
Alexei

_______________________________________________
Visit our blog: http://blog.pjsip.org

pjsip mailing list
pjsip@lists.pjsip.org
http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org

--
Best regards,
Alexei
mailto:alex2grad@gmail.com

NI
Nanang Izzuddin
Thu, Aug 18, 2016 8:06 AM

Hi Alexei,

Thanks for the confirmation on the patch.

Also just updated pj_dump_config() as suggested in SVN trunk r5421.

BR,
nanang

On Wed, Aug 17, 2016 at 10:17 PM, Alexei Gradinari alex2grad@gmail.com
wrote:

Hello Nanang,

Can confirm this patched fixed the issue.
Could you, please, add logging of the PJ_HAS_IPV6 in pj_dump_config.

Regards,
Alexei

Wednesday, August 17, 2016, 12:33:55 AM, you wrote:

Hi Alexei,

Finally have managed to reproduce the issue and test the patch here, the
patch seems to work, so just checked in it (with slight modification) to
SVN trunk for ticket https://trac.pjsip.org/repos/ticket/1953.

BR,
nanang

On Tue, Aug 16, 2016 at 9:15 PM, Alexei Gradinari alex2grad@gmail.com
wrote:
Hello Nanang,

I tested on Linux when kernel compiled without IPv6

CONFIG_IPV6 is not set

Regards,
Alexei

Tuesday, August 16, 2016, 2:45:58 AM, you wrote:

Hi Alexei,

Thank you for the report. The bug sounds valid, unfortunately I could not
reproduce the issue on Windows 10 and Linux, have tried to manually disable
IPv6 but still init_sock() did not fail (perhaps IPv6 loopback interface is
still up). So I am afraid we'd need some more help to test the attached
patch, thanks again in advance.

BR,
nanang

On Wed, Aug 10, 2016 at 4:20 AM, Alexei Gradinari alex2grad@gmail.com
wrote:
Hello,

The function init_sock at pjlib-util/src/pjlib-util/resolver.c
should check if the IPv6 is available instead of return an error.

Regards,
Alexei


Visit our blog: http://blog.pjsip.org

pjsip mailing list
pjsip@lists.pjsip.org
http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org

*-- Best regards, Alexei                            *
mailto:alex2grad@gmail.com alex2grad@gmail.com

Hi Alexei, Thanks for the confirmation on the patch. Also just updated pj_dump_config() as suggested in SVN trunk r5421. BR, nanang On Wed, Aug 17, 2016 at 10:17 PM, Alexei Gradinari <alex2grad@gmail.com> wrote: > Hello Nanang, > > Can confirm this patched fixed the issue. > Could you, please, add logging of the PJ_HAS_IPV6 in pj_dump_config. > > Regards, > Alexei > > > Wednesday, August 17, 2016, 12:33:55 AM, you wrote: > > Hi Alexei, > > Finally have managed to reproduce the issue and test the patch here, the > patch seems to work, so just checked in it (with slight modification) to > SVN trunk for ticket https://trac.pjsip.org/repos/ticket/1953. > > BR, > nanang > > > On Tue, Aug 16, 2016 at 9:15 PM, Alexei Gradinari <alex2grad@gmail.com> > wrote: > Hello Nanang, > > I tested on Linux when kernel compiled without IPv6 > # CONFIG_IPV6 is not set > > Regards, > Alexei > > > Tuesday, August 16, 2016, 2:45:58 AM, you wrote: > > Hi Alexei, > > Thank you for the report. The bug sounds valid, unfortunately I could not > reproduce the issue on Windows 10 and Linux, have tried to manually disable > IPv6 but still init_sock() did not fail (perhaps IPv6 loopback interface is > still up). So I am afraid we'd need some more help to test the attached > patch, thanks again in advance. > > BR, > nanang > > > On Wed, Aug 10, 2016 at 4:20 AM, Alexei Gradinari <alex2grad@gmail.com> > wrote: > Hello, > > The function init_sock at pjlib-util/src/pjlib-util/resolver.c > should check if the IPv6 is available instead of return an error. > > Regards, > Alexei > > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > > > > > > *-- Best regards, Alexei * > mailto:alex2grad@gmail.com <alex2grad@gmail.com> >