Re-enable privileged about content process for about:home / about:newtab / about:welcome
Categories
(Firefox :: New Tab Page, task, P3)
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox76 | --- | verified |
People
(Reporter: mconley, Assigned: mconley)
References
Details
Attachments
(2 files)
This bug will cover flipping the preference to enable the separate privileged about content process by default on Nightly.
After enabling on Nightly, we might file a follow-up bug to let the new privileged about content process ride the trains out to release.
| Assignee | ||
Comment 1•1 year ago
|
||
Pushed by mconley@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/cf4256249b8f Re-enable separate privileged about content process for about:home, about:newtab, about:welcome on Nightly. r=NeilDeakin
Backout by csabou@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/0893859124cc Backed out changeset cf4256249b8f for bc failures on browser_openTabAndSendCertInfo.js.
Comment 4•1 year ago
|
||
Failure log: https://treeherder.mozilla.org/logviewer.html#?job_id=290623449&repo=autoland
Backout link: https://hg.mozilla.org/integration/autoland/rev/0893859124cc0124b901844d76eb257081a9157b
[task 2020-02-26T19:02:37.398Z] 19:02:37 INFO - TEST-INFO | started process screenshot
[task 2020-02-26T19:02:37.464Z] 19:02:37 INFO - TEST-INFO | screenshot: exit 0
[task 2020-02-26T19:02:37.465Z] 19:02:37 INFO - Buffered messages logged at 19:01:07
[task 2020-02-26T19:02:37.465Z] 19:02:37 INFO - Entering test bound openFromPopUp
[task 2020-02-26T19:02:37.466Z] 19:02:37 INFO - Testing openFromPopUp
[task 2020-02-26T19:02:37.466Z] 19:02:37 INFO - Buffered messages finished
[task 2020-02-26T19:02:37.467Z] 19:02:37 INFO - TEST-UNEXPECTED-FAIL | toolkit/components/certviewer/tests/browser/browser_openTabAndSendCertInfo.js | Test timed out -
[task 2020-02-26T19:02:37.467Z] 19:02:37 INFO - GECKO(8196) | MEMORY STAT | vsize 2104346MB | vsizeMaxContiguous 66705798MB | residentFast 357MB | heapAllocated 103MB
[task 2020-02-26T19:02:37.467Z] 19:02:37 INFO - TEST-OK | toolkit/components/certviewer/tests/browser/browser_openTabAndSendCertInfo.js | took 90111ms
[task 2020-02-26T19:02:37.467Z] 19:02:37 INFO - GECKO(8196) | [Child 12192: Main Thread]: I/DocShellAndDOMWindowLeak ++DOCSHELL 0000018236F48800 == 1 [pid = 12192] [id = {1b46eb9a-84f6-426f-b1a4-4cf5ae91f950}]
[task 2020-02-26T19:02:37.467Z] 19:02:37 INFO - GECKO(8196) | [Child 12192: Main Thread]: I/DocShellAndDOMWindowLeak ++DOMWINDOW == 3 (000001823D44C090) [pid = 12192] [serial = 13] [outer = 0000000000000000]
[task 2020-02-26T19:02:37.467Z] 19:02:37 INFO - GECKO(8196) | [Child 12192: Main Thread]: I/DocShellAndDOMWindowLeak ++DOMWINDOW == 4 (0000018236FA5C00) [pid = 12192] [serial = 14] [outer = 000001823D44C090]
[task 2020-02-26T19:02:37.493Z] 19:02:37 INFO - GECKO(8196) | [Child 12192: Main Thread]: I/DocShellAndDOMWindowLeak ++DOMWINDOW == 5 (000001823D478000) [pid = 12192] [serial = 15] [outer = 000001823D44C090]
[task 2020-02-26T19:02:37.494Z] 19:02:37 INFO - checking window state
[task 2020-02-26T19:02:37.494Z] 19:02:37 INFO - Not taking screenshot here: see the one that was previously logged
[task 2020-02-26T19:02:37.494Z] 19:02:37 INFO - TEST-UNEXPECTED-FAIL | toolkit/components/certviewer/tests/browser/browser_openTabAndSendCertInfo.js | Found a after previous test timed out -
[task 2020-02-26T19:02:37.494Z] 19:02:37 INFO - Not taking screenshot here: see the one that was previously logged
[task 2020-02-26T19:02:37.494Z] 19:02:37 INFO - TEST-UNEXPECTED-FAIL | toolkit/components/certviewer/tests/browser/browser_openTabAndSendCertInfo.js | Found a browser window after previous test timed out -
[task 2020-02-26T19:02:37.494Z] 19:02:37 INFO - GECKO(8196) | must wait for focus
[task 2020-02-26T19:02:37.615Z] 19:02:37 INFO - GECKO(8196) | [GPU 9596, Compositor] WARNING: Possibly dropping task posted to updater thread: file z:/build/build/src/gfx/layers/apz/src/APZUpdater.cpp, line 428
[task 2020-02-26T19:02:37.625Z] 19:02:37 INFO - GECKO(8196) | [Parent 2052, Main Thread] WARNING: NS_ENSURE_TRUE(GetWrapper()) failed: file z:/build/build/src/dom/ipc/JSWindowActor.cpp, line 66
[task 2020-02-26T19:02:37.635Z] 19:02:37 INFO - TEST-START | toolkit/components/certviewer/tests/browser/browser_renderCertToUI.js
[task 2020-02-26T19:02:37.654Z] 19:02:37 INFO - GECKO(8196) | [Child 12968: Main Thread]: I/DocShellAndDOMWindowLeak ++DOCSHELL 000002BB19A4D000 == 4 [pid = 12968] [id = {19ff790a-a2d6-4748-abc6-593c2f86630b}]
[task 2020-02-26T19:02:37.654Z] 19:02:37 INFO - GECKO(8196) | [Child 12968: Main Thread]: I/DocShellAndDOMWindowLeak ++DOMWINDOW == 7 (000002BB200469A0) [pid = 12968] [serial = 106] [outer = 0000000000000000]
[task 2020-02-26T19:02:37.654Z] 19:02:37 INFO - GECKO(8196) | [Child 12968: Main Thread]: I/DocShellAndDOMWindowLeak ++DOMWINDOW == 8 (000002BB2007BC00) [pid = 12968] [serial = 107] [outer = 000002BB200469A0]
[task 2020-02-26T19:02:37.696Z] 19:02:37 INFO - GECKO(8196) | [Child 12968, Main Thread] WARNING: Trying to request nsIHttpChannel from DocumentChannel, this is likely broken: file z:/build/build/src/netwerk/ipc/DocumentChannel.cpp, line 64
[task 2020-02-26T19:02:37.696Z] 19:02:37 INFO - GECKO(8196) | [Child 12968: Main Thread]: I/DocShellAndDOMWindowLeak ++DOMWINDOW == 9 (000002BB22227800) [pid = 12968] [serial = 108] [outer = 000002BB200469A0]
[task 2020-02-26T19:02:37.701Z] 19:02:37 INFO - GECKO(8196) | [Parent 2052: Main Thread]: I/DocShellAndDOMWindowLeak --DOCSHELL 00000245FF5D8000 == 7 [pid = 2052] [id = {f10d7e22-ee81-42f7-96c2-29ee309d5792}] [url = about:blank]
[task 2020-02-26T19:02:38.693Z] 19:02:38 INFO - GECKO(8196) | MEMORY STAT | vsize 2104324MB | vsizeMaxContiguous 66705798MB | residentFast 356MB | heapAllocated 107MB
[task 2020-02-26T19:02:38.693Z] 19:02:38 INFO - TEST-OK | toolkit/components/certviewer/tests/browser/browser_renderCertToUI.js | took 1054ms
| Assignee | ||
Comment 5•1 year ago
|
||
The test was originally using BrowserTestUtils.waitForNewWindow with an argument for
the expected initial tab URL which was a function rather than a string. This helper
function never expects a function, but transparently passed it along to browserLoaded,
which can handle a function, so everything worked.
With the privileged about content process enabled, the waitForNewWindow code fell
down a codepath that doesn't handle the function parameter at all, and causes the
test to wait for a XULFrameLoaderCreated event that will never fire.
This patch adjusts the test to no longer pass the function to waitForNewWindow, since
it never supported having a function passed to it. Instead, we do the check for the
initial tab URL after the window has been opened.
| Assignee | ||
Updated•1 year ago
|
Pushed by mconley@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/623af81c196a Adjust browser_openTabAndSendCertInfo test for running about:certificate in the privileged about content process. r=NeilDeakin https://hg.mozilla.org/integration/autoland/rev/10c0ead96d08 Re-enable separate privileged about content process for about:home, about:newtab, about:welcome on Nightly. r=NeilDeakin
Comment 7•1 year ago
|
||
Backed out for bc failures on browser_google_behavior.js
Backout link: https://hg.mozilla.org/integration/autoland/rev/4f2f9bda72ac1c91576d2ca18d8323fe4104d99f
Log link: https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=291056951&repo=autoland&lineNumber=15260
| Assignee | ||
Comment 8•1 year ago
|
||
Looks like window shutdown leaks when Fission is enabled. Amazing.
Investigating, thanks!
| Assignee | ||
Comment 9•1 year ago
•
|
||
So what I've been able to figure out is that by calling focus on the search input element here: https://searchfox.org/mozilla-central/rev/91f6c02fcf4c16f78fdc4417f61f192688294066/browser/components/search/test/browser/browser_google_behavior.js#171
We ultimately cause the shutdown leak to occur.
Also, this seems to only be true since bug 1556627 landed. If I check out the previous commit, fad5300d47a8, and enable the privileged about content process, I don't get the leak anymore.
Comment 10•1 year ago
|
||
See bug 1614268.
| Assignee | ||
Comment 11•1 year ago
|
||
Thanks, Henri. I can confirm that by clearing out the active window before the test ends, I no longer see the leak.
| Assignee | ||
Comment 12•1 year ago
|
||
Looks like we'll need to figure out a solution to bug 1614268 before this can land.
Comment 13•1 year ago
|
||
putting this in the backlog as a P3 to remove from our traige queue - feel free to triage/reassign as you'd like, mconley! :D
| Assignee | ||
Comment 14•1 year ago
|
||
I'm just going to disable the leaking test for debug builds when Fission is enabled while I wait for bug 1614268 to land.
Try push: https://treeherder.mozilla.org/#/jobs?repo=try&revision=e8721bdc7b1782af1a67598c1ae1bc1607a1e156
Comment 15•1 year ago
|
||
Pushed by mconley@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/bce3596818fd Adjust browser_openTabAndSendCertInfo test for running about:certificate in the privileged about content process. r=NeilDeakin https://hg.mozilla.org/integration/autoland/rev/718652138136 Re-enable separate privileged about content process for about:home, about:newtab, about:welcome on Nightly. r=NeilDeakin
Comment 16•1 year ago
|
||
| bugherder | ||
https://hg.mozilla.org/mozilla-central/rev/bce3596818fd
https://hg.mozilla.org/mozilla-central/rev/718652138136
Comment 17•1 year ago
|
||
I'm not sure if there's a bug tracking the test landing, but now would be a good time to retry the test.
Comment 18•1 year ago
|
||
I have verified, that the browser.tabs.remote.separatePrivilegedContentProcess pref is now set to true in the "about:config" page.
Verified using the latest Firefox Nightly (76.0a1 Build ID - 20200402095145)installed on Windows 10 x64, Mac 10.15.4 and Ubuntu 18.04 x64.
| Assignee | ||
Comment 19•1 year ago
|
||
(In reply to Henri Sivonen (:hsivonen) from comment #17)
I'm not sure if there's a bug tracking the test landing, but now would be a good time to retry the test.
Which test?
Comment 20•1 year ago
|
||
(In reply to Mike Conley (:mconley) (:⚙️) from comment #19)
(In reply to Henri Sivonen (:hsivonen) from comment #17)
I'm not sure if there's a bug tracking the test landing, but now would be a good time to retry the test.
Which test?
The one referred to in comment 14.
Description
•