There seems to be a very clear regression in boot performance between 2/21 and today. The setup: - Two flame devices - Both updated to base image v18D_nightly [0] - One updated to latest mozilla-central-flame-kk using flash_pvt tool Flame 1: - OS: 3.0.0.0-prerelase - Platform Version: 38.0a1 - Build Ident: 20150212010213 - Git rev: d5a71ced Flame 2: - OS: 3.0.0.0-prerelase - Platform Version: 41.0a1 - Build Ident: 20150527010201 - Git rev: 8ca93673 Flame two displays lockscreen 3-4 seconds later. Steps to reproduce: - reflash two flames with v18D_nightly - update one to today's b2g/gaia - reboot Reproducible: always I'll attach a video in a few minutes

Video of the bootup: https://youtu.be/_oagaJeydEc

Is it bug 1151672 ? Or is it especially with bug 1151672 being fixed?

It's definitely with bug 1151672 fixed. The regression has been introduced on March 31st/April 1st, and fixed 6 days ago. The regression I see is between Feb and today, so bug 1151672 should not affect it.

According to boot time on Raptor, we have average 35s-40s. Do we agree this number and mark this bug as worksforme? [1] http://raptor.mozilla.org/#/dashboard/script/core.js?device=flame-kk&branch=master&memory=319&series=reboot.*

This case is too old. According to numbers is hard to be referred as regression perspective in 2.5, mark this bug as WONTFIX. Please reopen this if you have concern or any clue to find root cause.