[Webkit-unassigned] [Bug 184445] New: [GTK][WPE] Race condition when destroying webprocesses

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Tue Apr 10 04:08:58 PDT 2018


https://bugs.webkit.org/show_bug.cgi?id=184445

            Bug ID: 184445
           Summary: [GTK][WPE] Race condition when destroying webprocesses
           Product: WebKit
           Version: WebKit Nightly Build
          Hardware: Unspecified
                OS: Unspecified
            Status: NEW
          Severity: Normal
          Priority: P2
         Component: WebKit Gtk
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: magomez at igalia.com
                CC: bugs-noreply at webkitgtk.org

When a WebPageProxy gets destroyed, this is that ideally happens:

* WebPageProxy::close() is called:
  * Sends a WebPage::Close() message to the webprocess, so it can properly close its page before shutting down the process
  * Calls WebProcessProxy::removeWebPage() which, in the end, closes the IPC connection to the webprocess
* At the same time, when the webprocess executes WebPage::close() in response to the WebPage::Close() message
  * releases the page resources
  * at the end calls WebProcess::removeWebPage(), which causes the web process termination


But there's the possibility that the webprocess never processes the WebPage::Close() message. This happens when it detects that the IPC connection has been closed before processing the message. In that case, the webprocess exits without properly closing its webpages, causing problems on situations where a proper deinitialization is required (for example deinitializing a hardware decoder used by the media player).

We need to ensure that the WebPage instances owned by the WebProcess are always properly closed before exiting.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20180410/05d28ade/attachment-0002.html>


More information about the webkit-unassigned mailing list