Alice0775 White Reporter
	Description • 8 years ago

Build Identifier:
https://hg.mozilla.org/mozilla-central/rev/1a5b53a831e5a6c20de1b081c774feb3ff76756c
Mozilla/5.0 (Windows NT 10.0; WOW64; rv:51.0) Gecko/20100101 Firefox/51.0 ID:20160829030202

With e10s, it becomes 40% slow.
Without e10s, the performance degradation does not occur almost

Reproducible: always

Steps To reproduce:
1. Open http://www.vsynctester.com/
2. Moving mouse in a circular motion on top of the content area for a while.


Actual Results:
becomes 75fps -> 40fps

Expected Results:
Reduction should be small.

	Alice0775 White Reporter
	Comment 1 • 8 years ago

Regression window:
https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=7e2e9a6c0f0fe1f84b55987afcdc01fb16a124a3&tochange=3f4b6a6542ec6be9689ae265db6db4391c7d79d8

Suspect: Bug 1222662

	Jim Mathies [:jimm]
	Comment 2 • 8 years ago

Sounds like a buildup of cursor events we forward. I'm not seeing it on this site so I'd bet it shows up on slow hardware.

	Kartikaya Gupta (email:kats@mozilla.staktrace.com) Assignee
	Comment 3 • 8 years ago

This is not really a graphics bug, moving to Core::General which is where bug 1222662 lived.

	Andrew Overholt [:overholt]
	Comment 4 • 8 years ago

Can we get some telemetry on mouse event forwarding latency so we can decide how serious it is?

	Andrew Overholt [:overholt]
	Comment 5 • 8 years ago

Or wait, did you mean in our event queue or in some sort of OS event queue? (/me doesn't know how this stuff works in case it wasn't clear) If the former, maybe bug 1303296 could help?

	Alice0775 White Reporter
	Comment 6 • 8 years ago

I can reproduce the problem on 50.0.
However, I can no longer reproduce the problem on Nightly53.0a1 and 51.0b2.

Fixed range:
https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=9e9080a2e3462fc576b679603853f4395fadff8f&tochange=75abc730f8202005d6469a7ec324b7aba5a89374

Fixed by: 75abc730f820	Kartikaya Gupta — Bug 1300421 - Back out 4 csets from bug 1288760 for regressing event coordinate reporting. r=jfkthame

	Ritu Kothari (:ritu) (Inactive, please n-i to RyanVM, jcristau, or pascal)
	Comment 7 • 8 years ago

Too late to fix in 50.1.0 release