Eric, could this be related to some of the recent CLs that went into ImageLoader?

It is a double regression in this range
https://crosland.corp.google.com/log/9502.0.0..9505.0.0

Could this be responsible for ARM?
https://chromium-review.googlesource.com/#/c/490446/

That's mtpd, not ImageLoader, right?

Are you sure this is happening on ARM boards? It looks like all the failures in the wmatrix link are intel.

Also, there are no imageloader CLs in the range specified in c#2. The only CLs I put in relate to a debugd refactor which shouldn't have any effect on imageloader.

I'm not sure why any of the CLs in that range should have caused an issue, but everything does look pretty green before 9502...

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/platform/imageloader/+/d80e3eca9bbd0b0e9dbd26f39a018f02f224d45a

commit d80e3eca9bbd0b0e9dbd26f39a018f02f224d45a
Author: Greg Kerr <kerrnel@chromium.org>
Date: Tue May 02 02:18:18 2017

Initialize OpenSSL temporary pointer to nullptr.

Per https://www.openssl.org/docs/faq.html#PROG3, the temporary
variable passed to i2d_PUBKEY must be initialized to NULL,
or the OpenSSL API thinks it is a valid buffer that was
already allocated.

BUG= chromium:717030 
TEST=autotest platform_ImageLoaderServer

Change-Id: I77c668f2da3346a77be8c0efbbd3185a217de0fd
Reviewed-on: https://chromium-review.googlesource.com/491989
Commit-Ready: Greg Kerr <kerrnel@chromium.org>
Tested-by: Greg Kerr <kerrnel@chromium.org>
Reviewed-by: Eric Caruso <ejcaruso@chromium.org>
Reviewed-by: Mike Frysinger <vapier@chromium.org>

[modify] https://crrev.com/d80e3eca9bbd0b0e9dbd26f39a018f02f224d45a/imageloader_main.cc

Green again.
https://wmatrix.googleplex.com/unfiltered?hide_missing=True&releases=tot&tests=platform_ImageLoaderServer&days_back=20

Still not green. 
https://wmatrix.googleplex.com/unfiltered?hide_missing=True&releases=tot&tests=component_UpdateFlash&days_back=14

I may have to start investigating on the update flash side as this was a double regression.

reassigning while you look at this

Failed: download-omaha-component.

Something is going wrong there, pretty much every board is failing. What the heck.

I reproduced on my Samus. waffles@ thinks this is the bug fixed in https://codereview.chromium.org/2856573003/ and we should see the tests go green once Chrome rolls to the latest canary.

Waiting for 60.0.3087.0...

I need to add expectations for veyron_rialto and guado/moblab which don't have Flash. Otherwise green.