[asterisk-bugs] [JIRA] (ASTERISK-28463) Asterisk crash with SIGSEV due a bad PJSIP aor contact
Juan Martin (JIRA)
noreply at issues.asterisk.org
Thu Jun 27 06:22:47 CDT 2019
[ https://issues.asterisk.org/jira/browse/ASTERISK-28463?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Juan Martin updated ASTERISK-28463:
-----------------------------------
Description:
Hi,
I discovered that if you put a bad contact in the aor configuration, it crashes asterisk when the phone is registered, concretelly with segmentation fault (SIGSEGV).
I tested it with 16.2.1 and it's also reproducible with 16.4.0.
h5. Example:
\[200]
type=aor
max_contacts=1
{{contact=sip:200@*:5060}}
qualify_frequency=60
h6. If the phone is not connected it causes some errors but asterisk continue working:
{{\[2019-06-27 12:08:48.462] ERROR\[20447]: res_pjsip.c:3859 create_out_of_dialog_request: Unable to create outbound OPTIONS request to endpoint 200 as URI 'sip:200@*:5060' is not valid}}
{{\[2019-06-27 12:08:48.462] ERROR\[20447]: res_pjsip/pjsip_options.c:877 sip_options_qualify_contact: Unable to create request to qualify contact sip:200@*:5060 on AOR 200}}
h6. Then, as soon as the phone registers in asterisk it crashes the main process:
{{CLI> -- Added contact 'sip:200 at 192.168.75.102:5060' to AOR '200' with expiration of 3600 seconds}}
{{Segmentation fault (`core' generated)}}
h6. In syslog:
{{Jun 26 13:48:43 desarrolloV3 kernel: asterisk\[2285]: segfault at 0 ip 00007fa597ec11f8 sp 00007fa592c1eb30 error 4 in res_pjsip_path.so\[7fa597ec0000+3000]}}
{{Jun 26 13:48:43 desarrolloV3 kernel: Code: 8d 64 24 c8 48 8b 07 48 c7 45 c8 00 00 00 00 48 85 c0 74 05 80 38 00 75 11 48 8d 65 d8 31 c0 5b 41 5c 41 5d 41 5e 41 5f 5d c3 <48> 8b 06 48 89 fb 48 89 f7 ff 50 08 48 89 45 a8 49 89 c6 48 8b 40}}
{{Jun 26 13:48:43 desarrolloV3 systemd\[1]: asterisk.service: Main process exited, code=killed, status=11/SEGV}}
{{Jun 26 13:48:43 desarrolloV3 systemd\[1]: asterisk.service: Control process exited, code=exited status=1}}
h5. Conclusion
Avoiding the part of why there is a wildcard in the contact instead an IP address (contact=sip:200@*:5060), it's bad, I know. But I think that it should
not crash the entire system a bad line in configuration.
Write access to config files is required to exploit this problem. Severity: low.
Perhaps you could include a filter when parsing the pjsip config files.
If you remove the malformed contact line in the sample config, asterisk works fine.
I'll upload the backtraces to provide more info about the crash.
was:
Hi,
I discovered that if you put a bad contact in the aor configuration, it crashes asterisk when the phone is registered, concretelly with segmentation fault (SIGSEGV).
I tested it with 16.2.1 and it's also reproducible with 16.4.0.
h5. Example:
\[200]
type=aor
max_contacts=1
{{contact=sip:200@*:5060}}
qualify_frequency=60
h6. If the phone is not connected it causes some errors but asterisk continue working:
{{\[2019-06-27 12:08:48.462] ERROR\[20447]: res_pjsip.c:3859 create_out_of_dialog_request: Unable to create outbound OPTIONS request to endpoint 200 as URI 'sip:200@*:5060' is not valid}}
{{\[2019-06-27 12:08:48.462] ERROR\[20447]: res_pjsip/pjsip_options.c:877 sip_options_qualify_contact: Unable to create request to qualify contact sip:200@*:5060 on AOR 200}}
h6. Then, as soon as the phone registers in asterisk it crashes the main process:
{{CLI> -- Added contact 'sip:200 at 192.168.75.102:5060' to AOR '200' with expiration of 3600 seconds}}
{{Segmentation fault (`core' generated)}}
h6. In syslog:
{{Jun 26 13:48:43 desarrolloV3 kernel: asterisk\[2285]: segfault at 0 ip 00007fa597ec11f8 sp 00007fa592c1eb30 error 4 in res_pjsip_path.so\[7fa597ec0000+3000]}}
{{Jun 26 13:48:43 desarrolloV3 kernel: Code: 8d 64 24 c8 48 8b 07 48 c7 45 c8 00 00 00 00 48 85 c0 74 05 80 38 00 75 11 48 8d 65 d8 31 c0 5b 41 5c 41 5d 41 5e 41 5f 5d c3 <48> 8b 06 48 89 fb 48 89 f7 ff 50 08 48 89 45 a8 49 89 c6 48 8b 40}}
{{Jun 26 13:48:43 desarrolloV3 systemd\[1]: asterisk.service: Main process exited, code=killed, status=11/SEGV}}
{{Jun 26 13:48:43 desarrolloV3 systemd\[1]: asterisk.service: Control process exited, code=exited status=1}}
h5. Conclusion
Avoiding the part of why there is a wildcard in the contact instead an IP address (contact=sip:200@*:5060), it's bad, I know, But I think that it should
not crash the entire system a bad line in configuration.
Write access to config files is required to exploit this problem. Severity: low.
Perhaps you could include a filter when parsing the pjsip config files.
If you remove the malformec contact line in the sample config, asterisk works fine.
I'll upload the backtraces to provide more info about the crash.
> Asterisk crash with SIGSEV due a bad PJSIP aor contact
> ------------------------------------------------------
>
> Key: ASTERISK-28463
> URL: https://issues.asterisk.org/jira/browse/ASTERISK-28463
> Project: Asterisk
> Issue Type: Bug
> Security Level: None
> Components: Resources/res_pjsip_path
> Affects Versions: 16.2.1, 16.4.0
> Environment: Debian Stretch 9.9 (Intel x86_64)
> Reporter: Juan Martin
> Severity: Minor
> Labels: pjsip
> Attachments: backtrace_16.2.1.7z, backtrace_16.4.0.7z, pjsip_endpoints.conf
>
>
> Hi,
> I discovered that if you put a bad contact in the aor configuration, it crashes asterisk when the phone is registered, concretelly with segmentation fault (SIGSEGV).
> I tested it with 16.2.1 and it's also reproducible with 16.4.0.
> h5. Example:
> \[200]
> type=aor
> max_contacts=1
> {{contact=sip:200@*:5060}}
> qualify_frequency=60
> h6. If the phone is not connected it causes some errors but asterisk continue working:
> {{\[2019-06-27 12:08:48.462] ERROR\[20447]: res_pjsip.c:3859 create_out_of_dialog_request: Unable to create outbound OPTIONS request to endpoint 200 as URI 'sip:200@*:5060' is not valid}}
> {{\[2019-06-27 12:08:48.462] ERROR\[20447]: res_pjsip/pjsip_options.c:877 sip_options_qualify_contact: Unable to create request to qualify contact sip:200@*:5060 on AOR 200}}
> h6. Then, as soon as the phone registers in asterisk it crashes the main process:
> {{CLI> -- Added contact 'sip:200 at 192.168.75.102:5060' to AOR '200' with expiration of 3600 seconds}}
> {{Segmentation fault (`core' generated)}}
> h6. In syslog:
> {{Jun 26 13:48:43 desarrolloV3 kernel: asterisk\[2285]: segfault at 0 ip 00007fa597ec11f8 sp 00007fa592c1eb30 error 4 in res_pjsip_path.so\[7fa597ec0000+3000]}}
> {{Jun 26 13:48:43 desarrolloV3 kernel: Code: 8d 64 24 c8 48 8b 07 48 c7 45 c8 00 00 00 00 48 85 c0 74 05 80 38 00 75 11 48 8d 65 d8 31 c0 5b 41 5c 41 5d 41 5e 41 5f 5d c3 <48> 8b 06 48 89 fb 48 89 f7 ff 50 08 48 89 45 a8 49 89 c6 48 8b 40}}
> {{Jun 26 13:48:43 desarrolloV3 systemd\[1]: asterisk.service: Main process exited, code=killed, status=11/SEGV}}
> {{Jun 26 13:48:43 desarrolloV3 systemd\[1]: asterisk.service: Control process exited, code=exited status=1}}
> h5. Conclusion
> Avoiding the part of why there is a wildcard in the contact instead an IP address (contact=sip:200@*:5060), it's bad, I know. But I think that it should
> not crash the entire system a bad line in configuration.
> Write access to config files is required to exploit this problem. Severity: low.
> Perhaps you could include a filter when parsing the pjsip config files.
> If you remove the malformed contact line in the sample config, asterisk works fine.
> I'll upload the backtraces to provide more info about the crash.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
More information about the asterisk-bugs
mailing list