Bug 183075

Summary: Release assertion in WebPage::updatePreferences
Product: WebKit Reporter: Ryosuke Niwa <rniwa>
Component: WebKit2Assignee: Ryosuke Niwa <rniwa>
Status: RESOLVED FIXED    
Severity: Normal CC: ap, beidson, cdumez, commit-queue, mitz, youennf
Priority: P2 Keywords: InRadar
Version: WebKit Nightly Build   
Hardware: Unspecified   
OS: Unspecified   
Attachments:
Description Flags
Patch
none
Patch
none
Patch for landing
none
Patch for landing rniwa: commit-queue+

Ryosuke Niwa
Reported 2018-02-22 23:25:26 PST
After https://trac.webkit.org/changeset/228589, some people are hitting release assertions. There appears to be a race condition between UI process and WebContent process's preferences being synced vs web view being created. <rdar://problem/37790257>
Attachments
Patch (5.19 KB, patch)
2018-02-22 23:34 PST, Ryosuke Niwa 		no flags
DetailsFormatted DiffDiff
Patch (5.26 KB, patch)
2018-02-24 16:24 PST, Ryosuke Niwa 		no flags
DetailsFormatted DiffDiff
Patch for landing (4.93 KB, patch)
2018-02-26 11:33 PST, Ryosuke Niwa 		no flags
DetailsFormatted DiffDiff
Patch for landing (4.95 KB, patch)
2018-02-26 12:38 PST, Ryosuke Niwa 		rniwa: commit-queue+
DetailsFormatted DiffDiff
Show Obsolete (3) View All Add attachment proposed patch, testcase, etc.
Ryosuke Niwa
Comment 1 2018-02-22 23:34:07 PST
Created attachment 334506 [details] Patch
Ryosuke Niwa
Comment 2 2018-02-24 16:24:32 PST
Created attachment 334568 [details] Patch
youenn fablet
Comment 3 2018-02-26 10:43:37 PST
Comment on attachment 334568 [details] Patch r=me with non sw-enabled build fix. View in context: https://bugs.webkit.org/attachment.cgi?id=334568&action=review > Source/WebKit/StorageProcess/StorageProcess.cpp:191 > + return; The work done here seems harmless so maybe we can do it anyway, especially since we are disabling any IPC to Messages::WebSWServerToContextConnection? I am not familiar with this routine but maybe there is a chance for a race condition somehow so that we initialize the website data store first and get the entitlement later on.
Chris Dumez
Comment 4 2018-02-26 10:44:40 PST
*** This bug has been marked as a duplicate of bug 183135 ***
Chris Dumez
Comment 5 2018-02-26 10:49:15 PST
Comment on attachment 334568 [details] Patch lgtm too.
Ryosuke Niwa
Comment 6 2018-02-26 11:33:54 PST
Reopening to attach new patch.
Ryosuke Niwa
Comment 7 2018-02-26 11:33:55 PST
Created attachment 334631 [details] Patch for landing
Ryosuke Niwa
Comment 8 2018-02-26 11:34:10 PST
Comment on attachment 334631 [details] Patch for landing Wait for EWS.
Chris Dumez
Comment 9 2018-02-26 11:34:40 PST
*** Bug 183135 has been marked as a duplicate of this bug. ***
Ryosuke Niwa
Comment 10 2018-02-26 12:37:27 PST
Comment on attachment 334568 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=334568&action=review >> Source/WebKit/StorageProcess/StorageProcess.cpp:191 >> + return; > > The work done here seems harmless so maybe we can do it anyway, especially since we are disabling any IPC to Messages::WebSWServerToContextConnection? > I am not familiar with this routine but maybe there is a chance for a race condition somehow so that we initialize the website data store first and get the entitlement later on. We need this check to disable service worker in the storage process.
Ryosuke Niwa
Comment 11 2018-02-26 12:38:38 PST
Created attachment 334634 [details] Patch for landing
Ryosuke Niwa
Comment 12 2018-02-26 12:44:52 PST
Committed r229037: <https://trac.webkit.org/changeset/229037>
Note You need to log in before you can comment on or make changes to this bug.