1062899 - 2% win*/osx10.6 session restore regression on inbound (v.34) August 21 from bug 691601

Status: RESOLVED WONTFIX

(Testing :: Talos, defect)

Description

[Joel Maher ( :jmaher ) (UTC -8)]

session restore was hanging out <1210 for a long while, then it jumped up on August 21:
http://graphs.mozilla.org/graph.html#tests=[[313,63,37]]&sel=none&displayrange=30&datatype=running

here are some retriggers to highlight it:
https://tbpl.mozilla.org/?tree=Mozilla-Inbound&fromchange=0ed4b3b8d79b&tochange=c4d1c0041347&jobname=Windows%20XP%2032-bit%20mozilla-inbound%20pgo%20talos%20other

It appears that the patch from bug 952977 is where we regress:
https://hg.mozilla.org/integration/mozilla-inbound/rev/3325dc888b50

but we go even higher in the next push in the pushlog from bug 691601.

Comment 1

[Joel Maher ( :jmaher ) (UTC -8)]

Yoric, could you take a brief look at this and offer some advice

Comment 2

[David Zbarsky (:dzbarsky)]

That changeset from bug 952977 didn't touch any code that runs on Windows.

Comment 3

[David Teller [:Yoric] - still alive but not very active]

The problem is clearly not due to changes in Session Restore as that part of the code hasn't seen changes in a few months. Tim, any idea?

Comment 4

[Tim Taubert [:ttaubert] (inactive)]

I don't have a good idea off-hand. We indeed did not make any (bigger) changes to sessionstore in the last months. We consistently add stuff to the startup path though, initializing modules etc. That might also explain the slow rise here?

Comment 5

[Joel Maher ( :jmaher ) (UTC -8)]

this shows up on osx 10.6 as well.  I did some retriggers and bug 691601 clearly causes a regression:
https://tbpl.mozilla.org/?tree=Mozilla-Inbound&jobname=Rev4%20MacOSX%20Snow%20Leopard%2010.6%20mozilla-inbound%20talos%20other_nol64&fromchange=6ded38cc54e3&tochange=c4d1c0041347

Comment 6

[Joel Maher ( :jmaher ) (UTC -8)]

this also affects win7 and win8.

Comment 7

[Joel Maher ( :jmaher ) (UTC -8)]

we have already shipped this.