[Asterisk-code-review] thirdparty/pjproject: Fix deadlock in response retransmissions (asterisk[13])

Richard Mudgett asteriskteam at digium.com
Wed Jul 18 16:01:35 CDT 2018


Richard Mudgett has posted comments on this change. ( https://gerrit.asterisk.org/9477 )

Change subject: thirdparty/pjproject: Fix deadlock in response retransmissions
......................................................................


Patch Set 4:

There is a private clone_tdata() function in pjsip/src/pjsip-ua/sip_100rel.c that we'd need to use to clone tdata for similar things in a normal invite transaction to fix this deadlock.

Comment where clone_tdata() is called:
    /* Clone tdata.
     * We need to clone tdata because we may need to keep it in our
     * retransmission list, while the original dialog may modify it
     * if it wants to send another response.
     */


-> INVITE
<- 100 Trying
<- 183 Progress
<- 200 OK
<- 200 OK retransmission
-> ACK

Though a better question is why is a retransmission timer firing to re-send a 183 response?  That timer shouldn't even be started at this time in an INVITE transaction.  We would need to retransmit a final response if we don't receive an ACK.  However, we shouldn't be trying to send another response.


-- 
To view, visit https://gerrit.asterisk.org/9477
To unsubscribe, or for help writing mail filters, visit https://gerrit.asterisk.org/settings

Gerrit-Project: asterisk
Gerrit-Branch: 13
Gerrit-MessageType: comment
Gerrit-Change-Id: Ic381004a3a212fe1d8eca0e707fe09dba4a6ab4e
Gerrit-Change-Number: 9477
Gerrit-PatchSet: 4
Gerrit-Owner: Torrey Searle <tsearle at gmail.com>
Gerrit-Reviewer: George Joseph <gjoseph at digium.com>
Gerrit-Reviewer: Jenkins2
Gerrit-Reviewer: Richard Mudgett <rmudgett at digium.com>
Gerrit-Comment-Date: Wed, 18 Jul 2018 21:01:35 +0000
Gerrit-HasComments: No
Gerrit-HasLabels: No
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-code-review/attachments/20180718/bec71201/attachment.html>


More information about the asterisk-code-review mailing list