Perfherder has detected a talos performance regression from push 7cea08d78f8b23d7c8a41a76c9bb90c69df3a9bf. As author of one of the patches included in that push, we need your help to address this regression.

Regressions:

24% perf_reftest_singletons slow-selector-2.html windows7-32-shippable opt e10s stylo 0.06 -> 0.08
20% perf_reftest_singletons slow-selector-1.html windows7-32-shippable opt e10s stylo 0.07 -> 0.08
16% perf_reftest_singletons display-none-1.html windows7-32-shippable opt e10s stylo 0.07 -> 0.08

Details of the alert can be found in the alert summary, including links to graphs and comparisons for each of the affected tests. Please follow our guide to handling regression bugs and let us know your plans within 3 business days, or the offending patch(es) will be backed out in accordance with our regression policy.

For more information on performance sheriffing please see our FAQ.

The graph is quite noisy and this is unlikely to be Bug 1660798. I'll attempt to simulate a backout though to confirm.

Here is a try link with simulated backout so that we can check perf and determine if Bug 1660798 is related. https://treeherder.mozilla.org/#/jobs?repo=try&revision=16e51adf7ec4200c5cbee8abc800967bab974e07

Hopefully this gives a bit more clarity.

Here is a comparison simulating if we were landing Bug 1660798 today (ie "Base" has the code removed): https://treeherder.mozilla.org/perf.html#/comparesubtest?originalProject=try&newProject=try&newRevision=ff862be8b10fe35a5ca31cacf0b70e9f245637c5&originalSignature=1908963&newSignature=1908963&framework=1&originalRevision=16e51adf7ec4200c5cbee8abc800967bab974e07

I do see the the regressions but with a smaller percentage. The tests themselves are quite noisy and toggle between 0.06/0.08 so even a slight change might upset the equilibrium. The slow down itself is with the test and not in the CSS code it is testing. I think this issue is wont-fix unless we find regressions in areas more directly tied to JS perf.