814627 - WebRTC crash [@sipcc::PeerConnectionMedia::ice_media_stream]

Status: VERIFIED FIXED

(Core :: WebRTC, defect, P2)

Description

[Christoph Diehl [:posidron]]

Attachment: callstack

This crash is a rare one which happened till now only 1 time out of 1500 tests. 
Will look to get a testcase for it.

media/webrtc/signaling/src/peerconnection/PeerConnectionMedia.h:264
  mozilla::RefPtr<mozilla::NrIceMediaStream> ice_media_stream(size_t i) const {
    // TODO([e-mail]): If someone asks for a value that doesn't exist,
    // make one.
    if (i >= mIceStreams.size()) {

media/webrtc/signaling/src/media/VcmSIPCCBinding.cpp:559
  mozilla::RefPtr<NrIceMediaStream> stream = pc->impl()->media()->
    ice_media_stream(level-1);

media/webrtc/signaling/src/sipcc/core/gsm/lsm.c:648
          vcmRxAllocICE(media->cap_index, dcb->group_id, media->refid,
            lsm_get_ms_ui_call_handle(lcb->line, lcb->call_id, lcb->ui_id),
            dcb->peerconnection,
            media->level,
            &default_addr, &port_allocated,
            &candidates, &candidate_ct);

Tested with m-c changeset: 113925:20ec9014f220

Comment 1

[Randell Jesup [:jesup] (needinfo me)]

likely a null media() due to not having the Big Lock patch (ekr: sound right?)

Comment 2

[Jason Smith [:jsmith]]

If you manage to get a testcase for this that consistently reproduces the problem, feel free to change the in-testsuite flag in in-testsuite?

Comment 3

[Randell Jesup [:jesup] (needinfo me)]

Assigning to ekr, but on hold until Dispatch patch lands and we see if this recurs.  If we don't see it after that, we decide it's fixed, and reopen if it recurs.

Comment 5

[Christoph Diehl [:posidron]]

The crash did not occur again in my logs.

Tested with:
m-i changeset: 116243:9d6e95e77855
m-c changeset: 116160:c8a1314aa449

Comment 6

[Jason Smith [:jsmith]]

Verified per comment 5.