UI is intermittently non-responsive to any inputs (touchscreen, keyboard) |
|||
Issue descriptionChrome Version: 60.0.3112.41 beta OS: Chrome OS 9592.29.0 What steps will reproduce the problem? It seems not easily to find an efficient way to repro this issue but we start to note this issue recently on Elm devices at least. What is the expected result? UI is always responsive What happens instead? UI is intermittently non-responsive to any inputs Please use labels and text to provide additional information. Affected feedback reports: https://feedback.corp.google.com/#/Report/65732076718 https://listnr.corp.google.com/report/66953929034 https://listnr.corp.google.com/report/66504773823 https://listnr.corp.google.com/report/66506003539 https://listnr.corp.google.com/report/66580378582 Above feedback reports shows similar pattern below In the powerd.latest session seems like that at the time of the hang there was some "Audio activity started"/"Audio activity ongoing" followed by PushPullFIFO underflow/overflow message in CHROME_USER_LOG. It is hard to say whether audio pattern is a cause or an effect, but it seems to be some clues across these reports when UI/DUT was non-responsive. Jimmy, please help triage/route. Due to lacks of constant reproducible way, chrome://tracing is not easily deployed for clarifying why system was busy. I'm wondering if this could be a regression on R60 browser itself. Thanks
,
Jun 28 2017
The first one comes from 60.0.3112.26(Chrome OS: 9592.15.0), and the reaming four were reported at 60.0.3112.41 beta(Chrome OS: 9592.29.0) I haven't had a chance to check the same Chrome OS version on other devices yet. Few Caroline dogfooders stay on 9592.22.0, but I haven't noted this pattern from Caroline feedbak. May need to check Hana or Kevin to confirm whether its' Elm specific. Thanks
,
Jun 28 2017
The log in Chrome [1:19:0627/204645.823555:WARNING:PushPullFIFO.cpp(154)] PushPullFIFO: underflow while pulling (underflowCount=1, availableFrames=0, requestedFrames=512, fifoLength=8192) was saying that renderer did not prepare enough data. When system is busy due to whatever reason, renderer thread could not get enough cycle to do its job. That cause the PushPullFIFO error when browser thread wants to pull the data. Just like input is a victim, audio may as well be a victim. The first feedback report https://feedback.corp.google.com/product/208/neutron?lView=rd&lReport=65732076718 said there was reboot after hang. The crash id in the log leads to https://crash.corp.google.com/browse?stbtiq=bd1d45b7f0000000. The crash is related to (crbug 720078, crbug 720073, crbug 712382) But from https://bugs.chromium.org/p/chromium/issues/detail?id=720073#c8 , the crash should happen during shutdown, so this crash might be different issue. I have played with an elm EVT2 device on image 9625.0 using webaudio, youtube, webrtc for an hour and have not seen any issue. I think Daniel is a better owner than me since he has seen this issue the most, and there is no evidence that audio is the cause, and this is on elm. A handy way to check whether CRAS or Chrome consume all the CPU is to check "top" when issue happens.
,
Jun 30 2017
Looks like a browser hang was observed. No Audio pattern but a ton of shill/wpa_supplicant logging was observed on https://listnr.corp.google.com/report/66911855685 Description: Device became unresponsive with mouse clicks not being responsive and the mouse movement being jerky. Eventually, after 2-3 minutes, all my windows closed and then I could restore them. UI language: en-US Version: 60.0.3112.34 beta Product Specific Data (whitelisted): CHROME VERSION: 60.0.3112.34 beta CHROMEOS_RELEASE_DESCRIPTION: 9592.22.0 (Official Build) beta-channel caroline 2017-06-27T10:44:44.787733-07:00 ERR session_manager[1457]: [ERROR:object_proxy.cc(582)] Failed to call method: org.chromium.LibCrosServiceInterface.CheckLiveness: object_path= /org/chromium/LibCrosService: org.freedesktop.DBus.Error.NoReply: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken. 2017-06-27T10:45:00.249487-07:00 DEBUG kernel: [11692.403781] SELinux: initialized (dev proc, type proc), uses genfs_contexts 2017-06-27T10:45:19.752386-07:00 WARNING session_manager[1457]: [WARNING:liveness_checker_impl.cc(68)] Browser hang detected! 2017-06-27T10:45:19.752472-07:00 WARNING session_manager[1457]: [WARNING:liveness_checker_impl.cc(72)] Aborting browser process. 2017-06-27T10:45:19.752551-07:00 INFO session_manager[1457]: [INFO:browser_job.cc(157)] Terminating process: Browser did not respond to DBus liveness check. 2017-06-27T10:45:19.752573-07:00 INFO session_manager[1457]: [INFO:system_utils_impl.cc(110)] Sending 6 to 1483 as 1000 2017-06-27T10:45:22.812765-07:00 WARNING session_manager[1457]: [WARNING:browser_job.cc(165)] Aborting child process 1483's process group 3 seconds after sending signal 2017-06-27T10:45:22.812810-07:00 INFO session_manager[1457]: [INFO:browser_job.cc(149)] Terminating process group: Browser took more than 3 seconds to exit after signal. 2017-06-27T10:45:22.812834-07:00 INFO session_manager[1457]: [INFO:system_utils_impl.cc(110)] Sending 6 to -1483 as 1000 2017-06-27T10:45:24.991585-07:00 WARNING crash_reporter[23283]: Received crash notification for chrome[1483] user 1000 (called directly) 2017-06-27T10:45:24.991845-07:00 WARNING crash_reporter[23284]: [user] Received crash notification for chrome[1750] sig 6, user 1000 (ignoring call by kernel - chrome crash; waiting for chrome to call us directly) 2017-06-27T10:45:25.953792-07:00 WARNING crash_reporter[23313]: [user] Received crash notification for chrome[1483] sig 6, user 1000 (ignoring call by kernel - chrome crash; waiting for chrome to call us directly) 2017-06-27T10:45:25.953793-07:00 WARNING crash_reporter[23312]: Received crash notification for chrome-crash-unknown-process[23275] user 1000 (called directly) 2017-06-27T10:45:26.176341-07:00 WARNING crash_reporter[23318]: [user] Received crash notification for chrome[1819] sig 6, user 1000 (ignoring call by kernel - chrome crash; waiting for chrome to call us directly) 2017-06-27T10:45:26.195690-07:00 INFO session_manager[1457]: [INFO:child_exit_handler.cc(77)] Handling 1483 exit. 2017-06-27T10:45:26.195716-07:00 ERR session_manager[1457]: [ERROR:child_exit_handler.cc(85)] Exited with signal 6 2017-06-27T10:45:26.195784-07:00 INFO session_manager[1457]: [INFO:session_manager_service.cc(280)] Exiting process is chrome. 2017-06-27T10:45:26.195799-07:00 INFO session_manager[1457]: [INFO:browser_job.cc(149)] Terminating process group: Ensuring browser processes are gone.
,
Jul 3 2017
Two recent locked-up cases was observed on Caroline (60.0.3112.52, 9592.42.0) by the same reporter. https://listnr.corp.google.com/report/67434470523 https://listnr.corp.google.com/report/67357054259 The reporter should trigger alt + volup + x to recover DUT's frozen symptom manually so I checked into combined logs before that point. Nor OOM kills nor tab discards were found. From top section, it's did't point out too much clues.
,
Jul 3 2017
From problem description, two more instances were observed on Elm(60.0.3112.50, 9592.37.0) https://listnr.corp.google.com/report/67357352920 https://listnr.corp.google.com/report/67264900705
,
Jul 26 2017
A browser hang was detected on Elm in the following case (60.0.3112.50, 9592.37.0) https://listnr.corp.google.com/report/68518495289 I'm not sure what we need to figure this out and move forward. Seems like Caroline encountered the similar issue (browser hang) on crbug.com/737829 |
|||
►
Sign in to add a comment |
|||
Comment 1 by djkurtz@chromium.org
, Jun 28 2017Components: OS>Kernel>Audio
Labels: -Pri-2 OS-Chrome Pri-1
Status: Assigned (was: Untriaged)