Consistent chrome crash observed while using WebRTC - webrtc::RTCStatsCollector::ProduceMediaStreamAndTrackStats_s |
|||||||||||
Issue descriptionChrome Version: 65.0.3325.11 & 65.0.3325.9 OS: 10323.2.0 & 10323.1.0 What steps will reproduce the problem? (1) Login to Device (2) Join WebRTC (http://apprtc.appspot.com/?debug=loopback) (3) Now suspend/resume the device (either by closing the lid or powerd_dbus_suspend) What is the expected result? WebRTC should not crash What happens instead? Crash observed after resume Please use labels and text to provide additional information. Logs attached. Crashes: 2018-01-24T11:28:46.349192-08:00 NOTICE crash_sender[9473]: Crash report receipt ID f10d037917cc5b41 2018-01-24T11:33:05.647372-08:00 NOTICE crash_sender[11288]: Crash report receipt ID 956587eb0ff0756c 2018-01-24T11:33:08.679387-08:00 NOTICE crash_sender[11507]: Crash report receipt ID f698da9afaaea507 For graphics-related bugs, please copy/paste the contents of the about:gpu page at the end of this report.
,
Jan 24 2018
,
Jan 24 2018
Adding hta@ and hbos@ based on the blame list for the file.
,
Jan 25 2018
Hi, I can reproduce it as well. Uploaded Crash Report ID 4e03f197ac5d0fcf (Local Crash ID: ChromeOS) Crash report uploaded on Thursday, January 25, 2018 at 8:34:41 AM
,
Jan 25 2018
Comment 4: It was tested in Device Pyro with the version: 65.0.3325.9 / 10323.1.0 Additional info: It seem to reproduce in M64 as well. Version details are 64.0.3282.79 - Gnawty. Uploaded Crash Report ID d8cc8622cac177be (Local Crash ID: ChromeOS) Crash report uploaded on Thursday, January 25, 2018 at 9:05:31 AM
,
Feb 1 2018
Searching for crashes in this function. crash/9bf2352f30080dd8 (in 66 on Windows) seems to indicate that the problem is widespread and ongoing. Q for those who can reproduce: Does it reproduce on every suspend, or do you need to suspend/resume a couple of times to hit the bug?
,
Feb 1 2018
The crash in 9bf... points to line 1002 of rtcstatscollector.cc. The only version where line 1002 is within the ProduceMediaStreamAndTrackstats_s function is version 105 from 2018-1-24 - in which it points to the close bracket.
,
Feb 1 2018
@Comment 6: No it is straight forward. As soon as you suspend the laptop and resume you can observe it. Observed in Winky 65.0.3325.35 / 10323.9.0. Uploaded Crash Report ID 92e354416500b1d6 (Local Crash ID: ChromeOS) Crash report uploaded on Thursday, February 1, 2018 at 9:58:06 AM In Mac OS with stable M64 ->64.0.3282.119: It does seem a bit different. It blacks out the video and even audio is muted. I have to end the call and then proceed. It is not a crash. Is this behavior expected ?
,
Feb 1 2018
Hi all, Just to clarify, crash is happening to that particular tab and not the whole browser. As mentioned in the comment 8 (regarding macbook pro), just received a crash report. Even though the browser video and audio is muted, we do receive a crash. I hope it is useful. Crash details: Uploaded Crash Report ID 115ee471c4388f03 (Local Crash ID: 055985d5-2efe-4d74-a229-142af0c6fdfc) Crash report captured on Wednesday, January 31, 2018 at 1:57:04 PM, uploaded on Thursday, February 1, 2018 at 10:07:51 AM
,
Feb 1 2018
The Mac crasher is an OOM error. The crash in ProduceMediaStreamAndTrackStats_s *might* be a hidden OOM error too - the last thing that function does is to add the stats to a report, which is an object just created with new().
,
Feb 1 2018
Re: #6 - Crash is observed on every suspend and resume on that tab.
,
Feb 5 2018
Issue 805868 has been merged into this issue.
,
Feb 7 2018
Possibly related: https://crbug.com/807592
,
Feb 7 2018
Issue 807592 has been merged into this issue.
,
Feb 7 2018
Based on issue#807592 , tagging the bug as Beta blocker.
,
Feb 8 2018
,
Feb 8 2018
,
Feb 8 2018
hta@ should this be merged into https://crbug.com/807174 ?
,
Feb 8 2018
|
|||||||||||
►
Sign in to add a comment |
|||||||||||
Comment 1 by pgangishetty@chromium.org
, Jan 24 2018