1750742 - Enable win32k lockdown by default in Nightly

Status: RESOLVED FIXED

(Core :: Security: Process Sandboxing, task, P1)

Comment 1

[Chris Martin [:cmartin]]

Attachment: Bug 1750742 - Enable Win32k Lockdown by default in Nightly r=bobowen

It's time to graduate Win32k lockdown from Nightly Experiments to default on
Nightly.

Comment 2

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/052d3687f4f0
Enable Win32k Lockdown by default in Nightly r=bobowen

Comment 3

[Marian-Vasile Laza]

Backed out changeset 052d3687f4f0 (Bug 1750742) for causing marionette test failures.
Backout link
Push with failures
Failure Log

Comment 4

[Bob Owen (:bobowen)]

Try push with (hopefully) fixed enabled by default tests:
https://treeherder.mozilla.org/jobs?repo=try&revision=469b69c6bdddd8be968bfea038af1e2a6b284017

Comment 5

[Bob Owen (:bobowen)]

Attachment: Bug 1750742 p0: Fix test_win32k_enrollment.py for when win32k is enabled by default. r=tjr!

This also changes gen_win32k_tests.py to work on Windows.

Comment 6

[Bob Owen (:bobowen)]

(In reply to Bob Owen (:bobowen) from comment #4)

Try push with (hopefully) fixed enabled by default tests:
https://treeherder.mozilla.org/jobs?repo=try&revision=469b69c6bdddd8be968bfea038af1e2a6b284017

I've fixed the lint error (single quotes in gen script).

Comment 7

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/9ab08f172d80
p0: Fix test_win32k_enrollment.py for when win32k is enabled by default. r=tjr
https://hg.mozilla.org/integration/autoland/rev/cfc475962a9e
Enable Win32k Lockdown by default in Nightly r=bobowen

Comment 8

[Cristian Tuns]

https://hg.mozilla.org/mozilla-central/rev/9ab08f172d80
https://hg.mozilla.org/mozilla-central/rev/cfc475962a9e

Comment 9

[Alex Finder]

(In reply to Marian-Vasile Laza from comment #3)

Backed out changeset 052d3687f4f0 (Bug 1750742) for causing marionette test failures.
Backout link
Push with failures
Failure Log

== Change summary for alert #33418 (as of Tue, 01 Mar 2022 11:41:51 GMT) ==

Improvements:

Ratio	Test	Platform	Options	Absolute values (old vs new)
15%	cpstartup content-process-startup	windows10-64-shippable-qr	e10s fission stylo webrender-sw	69.46 -> 58.83
15%	cpstartup content-process-startup	windows10-64-shippable-qr	e10s fission stylo webrender	70.12 -> 59.92
11%	cpstartup content-process-startup	windows10-64-shippable-qr	e10s fission stylo webrender	67.07 -> 59.42

For up to date results, see: https://treeherder.mozilla.org/perfherder/alerts?id=33418

Comment 10

[Bob Owen (:bobowen)]

(In reply to Alex Finder from comment #9)

For clarity was it the back-out or the re-landing that caused the improvement?

Comment 11

[Alex Finder]

(In reply to Bob Owen (:bobowen) from comment #10)

(In reply to Alex Finder from comment #9)

For clarity was it the back-out or the re-landing that caused the improvement?

Between 9ab08f172d80 and cfc475962a9e, it was revision cfc475962a9e that caused the improvement (corresponding to patch D139487 ). Upon initial post, I thought it was related to the backout, but I can't find that revision in the backout link. So I suspect it's from the re-landing. Sorry for the confusion!