The following builds success for both boards. Maybe just a bad CL or someone already fixes this?

Another CQ round fails:

https://uberchromegw.corp.google.com/i/chromeos/builders/veyron_speedy-paladin/builds/4662
https://uberchromegw.corp.google.com/i/chromeos/builders/veyron_minnie-paladin/builds/1934

As xixuan@ says, it's plausible that this is caused by the CL chain ending at:

* https://chromium-review.googlesource.com/c/453411/

Specifically that touches kernel 3.14 and will return "segv" (aka sig 11) when the stack overflows.  The messages I see in the failing logs are:

--

2017-03-14T15:39:31.093534+00:00 WARNING crash_reporter[6027]: [user] Received crash notification for app_process32[5776] sig 11, user 655360 (ignoring - PID filtered out)
2017-03-14T15:39:31.094101+00:00 WARNING crash_reporter[6027]: [ARC] Received crash notification for zygote[5776] sig 11, user 655360 (developer build - not testing - always dumping)

--

I can't seem to find the actual information about the crash anywhere in autotest, so let me see if I can reproduce locally.

Yup, it was my CLs.  Thanks to smbarber@, tracked down to needing one more change:

* https://chromium-review.googlesource.com/455116
  UPSTREAM: mm: Don't count the stack guard page towards RLIMIT_STACK

Re-uploaded.  Sorry for not noticing that earlier.  :(

Yup, it was indeed my CL chain.  Simply logging in and starting up ARC++ was enough to reproduce.  Uploaded a new chain with the fix.