Closed Bug 1562036 Opened 3 months ago Closed 3 months ago

Re-enable "browser.tabs.remote.separatePrivilegedContentProcess" for AWSY tests

Categories

(Testing :: AWSY, defect, P1)

69 Branch
defect

Tracking

(firefox68 unaffected, firefox69 fixed)

RESOLVED FIXED
mozilla69
Tracking Status
firefox68 --- unaffected
firefox69 --- fixed

People

(Reporter: whimboo, Assigned: whimboo)

References

Details

Attachments

(1 file)

On bug 1558763 we had to disable this preference, but AWSY requires it. As such we have to re-enable it.

It's unclear to me which prefs files need to be updated. Is it prefs.json only, or also tp6-prefs.json? I assume the others aren't important here. Sadly there is no documentation what is used where in the tree.

Flags: needinfo?(erahm)
Blocks: 1561590
Depends on: 1558763
Version: 67 Branch → 69 Branch

In this case we need:

  • base-prefs.json
  • prefs.json
  • tp6-prefs.json
Flags: needinfo?(erahm)

(In reply to Henrik Skupin (:whimboo) [⌚️UTC+2] from comment #5)

Results are here:

https://treeherder.mozilla.org/perf.html#/compare?originalProject=mozilla-central&newProject=try&newRevision=df749c46e5c6b154a2d19c5d35c230f9000b3967&framework=4&selectedTimeRange=172800

But isn't even more worse now?

Yes, we know it's worse for the standard AWSY test due to adding an additional process. We accepted this regression previously (it also helps for base content process overhead, which matters for fission).

Flags: needinfo?(erahm)
Pushed by hskupin@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/a78ae8f1c287
[awsy] Re-enable "browser.tabs.remote.separatePrivilegedContentProcess". r=erahm
Status: ASSIGNED → RESOLVED
Closed: 3 months ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla69

== Change summary for alert #21687 (as of Mon, 01 Jul 2019 06:13:42 GMT) ==

Regressions:

16% Heap Unclassified linux64-shippable opt 70,441,706.42 -> 81,896,873.37
14% JS windows10-64-shippable opt stylo tp6 193,697,293.44 -> 221,617,441.37
13% Resident Memory windows7-32-shippable opt 540,540,770.19 -> 609,169,839.08
12% Heap Unclassified linux64-shippable opt stylo tp6 86,791,152.51 -> 97,237,532.30
12% Explicit Memory windows10-64-shippable opt stylo tp6 446,020,005.50 -> 499,290,888.51
12% Heap Unclassified macosx1014-64-shippable opt stylo tp6 113,930,175.18 -> 127,100,127.14
11% Heap Unclassified windows7-32-shippable opt stylo tp6 46,284,285.64 -> 51,436,416.69
11% Heap Unclassified macosx1014-64-shippable opt 107,358,944.36 -> 119,254,933.13
11% Heap Unclassified windows10-64-shippable opt stylo tp6 58,853,992.42 -> 65,142,299.00
10% JS windows10-64-shippable opt 106,090,976.73 -> 117,098,307.97
10% Explicit Memory linux64-shippable opt 350,496,396.34 -> 386,494,295.83
10% Resident Memory windows10-64-shippable opt stylo tp6 644,655,032.97 -> 710,273,338.49
10% JS windows10-64-shippable-qr opt 106,029,609.60 -> 116,799,313.71
10% Heap Unclassified windows10-64-shippable opt 43,075,751.22 -> 47,439,190.25
10% Explicit Memory windows10-64-shippable opt 321,212,826.33 -> 353,395,184.34
10% JS linux64-shippable opt 103,804,207.61 -> 114,084,987.33
10% Heap Unclassified windows7-32-shippable opt 33,192,924.27 -> 36,399,109.33
9% JS macosx1014-64-shippable opt 106,060,722.11 -> 115,638,951.14
9% JS linux64-shippable-qr opt 104,527,230.24 -> 113,908,682.15
9% JS windows7-32-shippable opt 78,051,317.75 -> 85,019,155.81
9% Explicit Memory macosx1014-64-shippable opt stylo tp6 517,986,921.49 -> 562,966,844.69
9% JS windows7-32-shippable opt stylo tp6 155,951,125.09 -> 169,297,098.34
9% Explicit Memory windows10-64-shippable-qr opt 336,012,740.34 -> 364,717,298.05
8% Explicit Memory windows7-32-shippable opt stylo tp6 373,421,858.01 -> 404,960,162.62
8% Explicit Memory linux64-shippable opt stylo tp6 494,287,656.01 -> 535,447,471.17
8% Resident Memory windows7-32-shippable opt stylo tp6 659,677,419.92 -> 713,246,933.60
8% Resident Memory windows10-64-shippable opt 501,055,889.79 -> 541,225,048.50
8% JS macosx1014-64-shippable opt stylo tp6 208,399,221.69 -> 225,092,861.65
8% Explicit Memory macosx1014-64-shippable opt 388,993,827.01 -> 419,960,919.08
8% Explicit Memory windows7-32-shippable opt 260,013,835.11 -> 280,531,961.46
8% JS linux64-shippable-qr opt stylo tp6 206,047,077.98 -> 222,281,545.18
8% Explicit Memory windows10-64-shippable-qr opt stylo tp6 486,847,946.36 -> 524,614,130.51
8% Resident Memory windows7-32-shippable opt stylo tp6 660,806,060.75 -> 711,897,873.16
7% JS linux64-shippable opt stylo tp6 205,897,253.15 -> 221,315,739.49
7% Heap Unclassified windows10-64-shippable-qr opt 51,368,036.01 -> 55,175,039.81
7% Resident Memory linux64-shippable opt 588,002,303.39 -> 630,513,679.12
7% Heap Unclassified windows10-64-shippable-qr opt stylo tp6 66,950,980.62 -> 71,709,367.19
7% Explicit Memory linux64-shippable-qr opt stylo tp6 596,591,078.67 -> 636,874,188.32
7% Resident Memory windows10-64-shippable-qr opt 528,493,436.30 -> 563,317,139.58
6% Resident Memory macosx1014-64-shippable opt stylo tp6 678,183,738.99 -> 718,613,817.48
5% Resident Memory linux64-shippable opt stylo tp6 748,671,238.50 -> 789,572,540.39
5% Resident Memory linux64-shippable-qr opt stylo tp6 891,849,232.42 -> 934,480,028.25

Improvements:

12% Images linux64-shippable-qr opt 6,700,126.00 -> 5,897,911.60
11% Images macosx1014-64-shippable opt 5,089,808.33 -> 4,520,823.70
8% Images windows10-64-shippable-qr opt 7,983,322.80 -> 7,367,808.65

For up to date results, see: https://treeherder.mozilla.org/perf.html#/alerts?id=21687

Usually for regressions we log a separate bug but as this is a framework change I will add a comment here.

:whimboo please confirm these regressions are expected

Flags: needinfo?(hskupin)

Parts of those are expected and should actually be the exact opposite of those changes from bug 1561590 comment 0. But why are there way lesser improvements logged compared to the regressions here? Did we miss to add all of them to the other bug?

Flags: needinfo?(hskupin)
Flags: needinfo?(fstrugariu)
Flags: needinfo?(erahm)

If you compare alerts:
https://treeherder.mozilla.org/perf.html#/alerts?id=21576
and
https://treeherder.mozilla.org/perf.html#/alerts?id=21687

You can see they are mirrored. We might have not detected some of them in the other bug. but the numbers are comparable

Flags: needinfo?(fstrugariu)

Ok, so I assume it's all good then? That's what we actually wanted to have.

Flags: needinfo?(erahm)

Henrik, yes, it's all good.

You need to log in before you can comment on or make changes to this bug.