WebKit Bugzilla
New
Browse
Search+
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
RESOLVED FIXED
88449
Worker tear-down can re-enter JSC during GC finalization
https://bugs.webkit.org/show_bug.cgi?id=88449
Summary
Worker tear-down can re-enter JSC during GC finalization
Mark Hahnenberg
Reported
2012-06-06 13:23:52 PDT
~AbstractWorker can enter JS during GC finalization due to attempting to update the Web Inspector when the worker is being torn down, which is not allowed by JSC. ~Worker has a similar story. Both of these should be fixed to prevent this.
Attachments
Patch
(2.25 KB, patch)
2012-06-06 14:53 PDT
,
Mark Hahnenberg
no flags
Details
Formatted Diff
Diff
Patch
(1.23 KB, patch)
2012-06-06 18:01 PDT
,
Mark Hahnenberg
no flags
Details
Formatted Diff
Diff
Show Obsolete
(1)
View All
Add attachment
proposed patch, testcase, etc.
Mark Hahnenberg
Comment 1
2012-06-06 14:53:30 PDT
Created
attachment 146116
[details]
Patch
Mark Hahnenberg
Comment 2
2012-06-06 15:06:32 PDT
> ~AbstractWorker can enter JS during GC finalization due to attempting to update the Web Inspector when the worker is being torn down, which is not allowed by JSC.
To make this more clear: ~AbstractWorker can re-enter JS during GC finalization, which is not allowed by JSC, due to attempting to update the Web Inspector when the worker is being torn down.
Geoffrey Garen
Comment 3
2012-06-06 15:21:17 PDT
Comment on
attachment 146116
[details]
Patch r=me
Mark Hahnenberg
Comment 4
2012-06-06 15:29:07 PDT
Committed
r119624
: <
http://trac.webkit.org/changeset/119624
>
Ryosuke Niwa
Comment 5
2012-06-06 16:59:40 PDT
A bunch of worker tests started crashing after this patch:
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fast%2Fworkers%2Fworker-event-listener.html%2Cfast%2Fworkers%2Fstorage%2Fchange-version-handle-reuse-worker.html%2Chttp%2Ftests%2Fhistory%2Fback-during-onload-triggered-by-back.html%2Cfast%2Fworkers%2Fstorage%2Fread-and-write-transactions-dont-run-together.html
WebKit Review Bot
Comment 6
2012-06-06 17:04:40 PDT
Re-opened since this is blocked by 88472
Mark Hahnenberg
Comment 7
2012-06-06 18:01:39 PDT
Created
attachment 146165
[details]
Patch
Mark Hahnenberg
Comment 8
2012-06-06 18:02:11 PDT
Let's try landing each piece of the patch separately to determine which half caused the regression. I think this first patch is the safer of the two.
Geoffrey Garen
Comment 9
2012-06-07 11:11:18 PDT
Comment on
attachment 146165
[details]
Patch r=me
WebKit Review Bot
Comment 10
2012-06-07 11:32:08 PDT
Comment on
attachment 146165
[details]
Patch Clearing flags on attachment: 146165 Committed
r119740
: <
http://trac.webkit.org/changeset/119740
>
WebKit Review Bot
Comment 11
2012-06-07 11:32:13 PDT
All reviewed patches have been landed. Closing bug.
Note
You need to
log in
before you can comment on or make changes to this bug.
Top of Page
Format For Printing
XML
Clone This Bug