ubuntu-webapps-bugs team mailing list archive
-
ubuntu-webapps-bugs team
-
Mailing list archive
-
Message #00394
[Bug 1304648] Re: Forks browser process to run sandbox IPC helper
This bug was fixed in the package oxide-qt - 1.0.0~bzr490-0ubuntu1
---------------
oxide-qt (1.0.0~bzr490-0ubuntu1) trusty; urgency=medium
* Update to r490
- Build with enable_plugins: 0 and toolkit_views: 0 to trim the size
a bit
- Ensure subframe navigations that require a new window become top-level
navigations for webviews that don't implement WebView.newViewRequested
- Add an option for enabling proprietary codecs (ENABLE_PROPRIETARY_CODECS)
- Build and install the l10n pak files
- Add file picker support (LP: #1260008)
- Add some resources to oxide.pak that were missing (eg, directory listing
HTML)
- Fix a crash that occurs when handling events from some mouse buttons
- Add cursor support (LP: #1257662)
- Reimplement Chromium's RenderSandboxHostLinux so that it runs the sandbox
IPC helper process as a proper child process rather than just forking
the browser process, which is dangerous for Oxide (LP: #1304648)
* Split libffmpegsumo.so in to 2 separate packages (oxideqt-codecs and
oxideqt-codecs-extra) (LP: #1301341)
-- Chris Coulson <chris.coulson@xxxxxxxxxxxxx> Tue, 08 Apr 2014 15:27:09 +0100
** Changed in: oxide-qt (Ubuntu)
Status: New => Fix Released
--
You received this bug notification because you are a member of Ubuntu
WebApps bug tracking, which is subscribed to Oxide.
https://bugs.launchpad.net/bugs/1304648
Title:
Forks browser process to run sandbox IPC helper
Status in Oxide Webview:
Fix Released
Status in “oxide-qt” package in Ubuntu:
Fix Released
Bug description:
Chromium forks the browser process to run a sandbox IPC helper
process, which provides a mechanism for sandboxed renderers to access
things like the font cache. However, it does this without execing a
new process image in the child.
Whilst this is fine for Chromium (because it forks at startup and
before any threads are created), it is bad for us for a few reasons:
By the time we "start" Chromium:
- We already have a QML app and Qt gunk running in the browser process - I suspect that a lot of the memory used by the app will get copied eventually, which is a waste.
- We already have many threads running. As threads don't get forked, the child process will deadlock if any thread held a lock it requires at some point.
To manage notifications about this bug go to:
https://bugs.launchpad.net/oxide/+bug/1304648/+subscriptions
References