[H264] Fatal errors on device Blaze with hardware acceleration video decode disabled |
||||||||||||
Issue descriptionVersion: 52.0.2743.28 / 8350.19.0 dev Device: Blaze Reference bug: https://bugs.chromium.org/p/chromium/issues/detail?id=615272#c25 What steps will reproduce the problem? (1) Join an apprtc loopback call from device Blaze using https://appr.tc/?debug=loopback&vsc=h264 (2) Stay in the call for 1-2 minutes What is the expected output? Audio and video should be clear without any freezes What do you see instead? 1) Remote video starts after 30 sec -1min after loopback call is started 2) Remote video freezes as soon as it starts 3) V4L2 Decode WeakPTr errors as shown in : https://paste.googleplex.com/6523866897711104 Note that this issue is not seen when "Hardware-accelarated video decode Mac, Windows, chrome OS" flag is disabled in chrome://flags
,
Jun 7 2016
,
Jun 7 2016
,
Jun 7 2016
,
Jun 7 2016
Moving this nonessential bug to the next milestone. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Jun 7 2016
Should WebRTC H264 related issues block M52?
,
Jun 8 2016
The above version seems not to have the revert in issue 615272 merged in yet. I'm also seeing evidence of HW encoder being invoked in the attached logs. Would you be able to retest with CrOS 8350.21.0 / 52.0.2743.32 please?
,
Jun 8 2016
posciak@, I think this bug is with SW encode and HW decode, the logs point to that. I am looking forward to srcv@ to reproduce it.
,
Jun 8 2016
This issue is not reproducible with apprtc loopback calls on device Blaze with 52.0.2743.32 / 8350.21.0 dev (tried with both Hardware decode enabled and disabled)
,
Jun 8 2016
Actually this is more likely to have been "fixed" by turning off HW enc.
,
Jun 8 2016
@8: FTR, the logs in #1 show evidence of HW encoder being invoked, so I believe this was actually tested with HW encoder still on.
,
Jun 8 2016
(The above refers to the original report, HW encoder was disabled for tests in #9 on 52.0.2743.32 / 8350.21.0 dev)
,
Jun 8 2016
Ok, sounds good then. I wasn't sure if disabling HW encode was enough as WeakPtr seemed to point at VDA.
,
Jun 20 2016
Marking this issue as "Verified" as per comment #10 |
||||||||||||
►
Sign in to add a comment |
||||||||||||
Comment 1 by srcv@chromium.org
, Jun 6 2016