New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 836120 link

Starred by 2 users

Issue metadata

Status: WontFix
Owner: ----
Closed: Oct 5
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Android
Pri: 2
Type: Bug



Sign in to add a comment

SkMessageBus will cause memory continue to grow and never release under some situations

Reported by roger...@gmail.com, Apr 24 2018

Issue description

Steps to reproduce the problem:
1. Open url - res.mayi.aiyounet.cn/demo6/index_alipay2.html, it is a 2d canvas game
2. Open a console and type "watch 'adb shell dumpsys meminfo com.android.chrome:sandboxed_process0'", watch the memory usage
3. Native Heap will continue to grow

What is the expected behavior?
Native Heap will not continue to grow

What went wrong?
Worker context and shared main thread context (render_thread_impl) will create two GrContext, and two GrResourceCache, and one SkMessageBus<GrUniqueKeyInvalidatedMessage> with two Inboxs, the canvas 2d game will continue to put messages into these two Inboxs but the Inbox belong to Worker context never get poll, and its memory will continue to grow and do not have the chance to release.

Did this work before? N/A 

Chrome version: 66.0.3359.126  Channel: n/a
OS Version: 7.0
Flash Version: Shockwave Flash 29.0 r0
 

Comment 1 by roger...@gmail.com, Apr 24 2018

Actually I found this bug in our customized WebView based on m57, but Chrome 66 also appear the same phenomenon.

Currently, my temporarily patch is to force worker context's GrResourceCache to purgeAsNeeded at the end of each Canvas2DLayerBridge::flush, so we can poll the SkMessageBus<GrUniqueKeyInvalidatedMessage>::Inbox of worker context to release the memory.
Labels: Needs-triage-Mobile
Components: Internals>Core
Labels: Triaged-Mobile TE-NeedsTriageHelp
This issue seems to be out of TE-scope, Hence adding TE-NeedsTriageHelp label for further triaging of the issue.

Thanks!!
Cc: nyerramilli@chromium.org ligim...@chromium.org pnangunoori@chromium.org sandeepkumars@chromium.org
Labels: -TE-NeedsTriageHelp -Triaged-Mobile
Sandeep, please follow the instruction for remote debugging and try reproducing the bug.

https://developers.google.com/web/tools/chrome-devtools/remote-debugging/
Labels: Needs-Feedback Triaged-Mobile
Tested the issue using #66.0.3359.126 on Samsung J7, 7.0.0 and didn't observed such behavior as per the steps mentioned below.

Steps:
1. Launched Browser
2. Navigated to res.mayi.aiyounet.cn/demo6/index_alipay2.html, it is a 2d canvas game in Mobile Chrome
3. Connected as remote devcie
4. Opened Console in Mac and typed "watch 'adb shell dumpsys meminfo com.android.chrome:sandboxed_process0'"
5. Navigated to Memory tab
6. No such behavior is observed

@roger2yi: Could you please find the above observations and help us with a screencast for better understanding of the issue?

Thanks!!

Comment 6 by roger...@gmail.com, May 7 2018

Need to stay at the page as screenshot show, and keep watch the memory usage at host
2018-05-07 14-19-50屏幕截图.png
116 KB View Download
2018-05-07 14-18-37屏幕截图.png
450 KB View Download
Project Member

Comment 7 by sheriffbot@chromium.org, May 7 2018

Labels: -Needs-Feedback
Thank you for providing more feedback. Adding the requester to the cc list.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Comment 8 by roger...@gmail.com, May 7 2018

After 10 ~ 20 mins
2018-05-07 14-43-37屏幕截图.png
114 KB View Download

Comment 9 by roger...@gmail.com, May 7 2018

After 20+ mins, native heap reach 400m+ and browser tab is crashed
2018-05-07 14-52-26屏幕截图.png
60.0 KB View Download
Labels: Needs-Feedback
@roger2yi: Can you please help us with a sample Crash I'd from chrome://crashes?

Thanks!!
Cc: chelamcherla@chromium.org
Status: WontFix (was: Unconfirmed)
Closing this issue due to lack of feedback from reporter. Please feel free to raise a new issue if this is still seen.

Thanks!

Sign in to add a comment