Identified (hopefully) the candidate by manual `git bisect`. Reverted https://chromium-review.googlesource.com/c/chromium/src/+/990314.

arthursonzogni@ assigning to you as you are the author of crrev.com/c/990314. If I have incorrectly identified your CL as the actual cause please assign this bug to me and accept my apologies for reverting your change.

Thanks Chris for tracking this down. I'll take a look and see if there's a quick fix, as this is a test-only issue and a bunch of work depends on this launching.

I have a fix for the CFI issue.

The LSAN bot is a different bug. It also started failing at 547259, while Arthur's patch is at 547350

(I'll try to bisect the lsan issue now)

and Chris already fixed the lsan issue by reverting https://chromium.googlesource.com/chromium/src/+/11df4514ddd778370c2316a3acb1b6da772d5291 :)

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/39d5daaa736c7a127121b6b91dc1254177948a7b

commit 39d5daaa736c7a127121b6b91dc1254177948a7b
Author: John Abd-El-Malek <jam@chromium.org>
Date: Mon Apr 02 21:31:41 2018

Fix bad cast from ResourceContext to ProfileIOData in ChromeContentBrowserClient.

In unit tests, this isn't true as the object is a MockResourceContext. For now the quick fix is to check for the safe browsing service being available before trying to add its throttle, as it isn't set in unit tests.

Bug:  828062 
Change-Id: Ibb2d65307b72f02fe1e0de27797ca834f01cb184
Reviewed-on: https://chromium-review.googlesource.com/990724
Reviewed-by: Nasko Oskov <nasko@chromium.org>
Commit-Queue: John Abd-El-Malek <jam@chromium.org>
Cr-Commit-Position: refs/heads/master@{#547522}
[modify] https://crrev.com/39d5daaa736c7a127121b6b91dc1254177948a7b/chrome/browser/chrome_content_browser_client.cc