Very high wakeup event in Chrome Helper
Reported by
proffrns...@gmail.com,
Jun 15 2016
|
||||||||||||||||||||||||||||||||
Issue descriptionUserAgent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/52.0.2743.33 Safari/537.36 Steps to reproduce the problem: 1. Use Google Chrome on macOS 10.12 2. 3. What is the expected behavior? What went wrong? Getting diag reports in Console with the following message from Google Chrome Helper: "Wakeups: 45005 wakeups over the last 222 seconds (203 wakeups per second average), exceeding limit of 150 wakeups per second over 300 seconds" Attached all logs relating to this issues. Did this work before? Yes OS X 10.11.5 Chrome version: 52.0.2743.33 Channel: beta OS Version: OS X 10.12.0 Flash Version: Shockwave Flash 22.0 r0
,
Jun 15 2016
See also Issue 619981 .
,
Jun 16 2016
,
Jun 16 2016
@proffrnswrth Are you willing to try using tomorrow's canary build? We had at least one 10.12-related fix committed today, although that addressed memory usage and not wakeups.
,
Jun 16 2016
Sure thing! Let me know when it goes live and I will check it out.
,
Jun 17 2016
,
Jun 18 2016
https://codereview.chromium.org/2075953002/ is the CL with the fix, but it has not landed yet.
,
Jun 21 2016
Assigning this tapted since he wrote the CL with the fix.
,
Jul 7 2016
The fix for Issue 619981 landed in r402523 and was merged to m52 Beta for version 52.0.2743.60. proffrnswrth@: do you still see this in 52.0.2743.60? Console.app currently always crashes on startup for me in the Sierra Beta 1 (16A201w). Beta 2 just out, so I'll give that a go. I don't have anything lying around in ~/Library/Logs/DiagnosticReports, but that was the case before r402523 landed as well. (Console.app crash. Probably not related to anything Chrome is doing. Crashed Thread: 2 Dispatch queue: All Messages :: NSOperation 0x6000000dbc10 (QOS: USER_INTERACTIVE) Exception Type: EXC_BAD_ACCESS (SIGSEGV) Exception Codes: EXC_I386_GPFLT Exception Note: EXC_CORPSE_NOTIFY Thread 0:: Dispatch queue: com.apple.main-thread 0 libobjc.A.dylib 0x00007fff9eeeeafe search_method_list(method_list_t const*, objc_selector*) + 4 1 libobjc.A.dylib 0x00007fff9eef0c82 lookUpImpOrForward + 240 2 libobjc.A.dylib 0x00007fff9eef06d4 _objc_msgSend_uncached + 68 3 com.apple.AppKit 0x00007fff8993680a __RepSelectionCandidateLayoutDirectionScore_block_invoke + 205 4 com.apple.AppKit 0x00007fff89936732 RepSelectionCandidateLayoutDirectionScore + 85 5 com.apple.AppKit 0x00007fff8993668b RepSelectionCompareLayoutDirection + 24 6 com.apple.AppKit 0x00007fff894b24b7 -[NSImage bestRepresentationAmongRepresentations:forRect:context:hints:] + 1253
,
Jul 7 2016
tapted@: I am still getting this in 54.0.2789.0 (2789.0). I am attaching new logs relevant to this build.
,
Jul 12 2016
I'm not sure where to go from here. Maybe robert or mark know better how to interpret these "excessive wakeups" diagnostic reports. It's also come up (not in Sierra) in Issue 368774. For these diagnostic reports, it seems to be a similar story. Chrome's activity being captured in the sample doesn't seem remarkable in any way. On my machine, the `kernel_task` process always ranks highest for `Idle Wake Ups` - perhaps that process is causing the excess, but can't be sampled by the OS watchdog, so just gets omitted in the report. Then Chrome is next, but far, far down the list of wakeup counts. Feeding the previously attached reports into go/crsym give an "Unknown Report Version 19", but I managed to get a symbolized report from a Chromium build (attached). Nothing really stands out. This report appeared in /Library/Logs/DiagnosticReports after watching videos on youtube.com for a bit, then pausing the video. I don't think anything in particular triggered it. Currently, I'm inclined to treat this as an OS bug, but if something Chrome is doing is tickling it a lot, that's something we may still need to address.
,
Aug 16 2016
The "Unknown Report Version 19" message can be fixed by manually changing the report version to 18. I started symbolizing several reports but nothing stands out. I see some audio decode, and flash (presumably for video decode). I see some extension activity. "203 wakeups per second average" is not ridiculously high for 60fps video, although ideally we'd be better. I tried playing a video on youtube and watching the console, but wasn't able to trigger this particular error.
,
Aug 16 2016
Without further information, I'm not sure what we can do about this. proff: Can you provide more detailed instructions about how to reproduce this problem? e.g. is there a particular video that you're watching?
,
Sep 13 2016
This bug has been without needed feedback for some time. Marking as WontFix. |
||||||||||||||||||||||||||||||||
►
Sign in to add a comment |
||||||||||||||||||||||||||||||||
Comment 1 by meh...@chromium.org
, Jun 15 2016Labels: Hotlist-Sierra