It looks like this was caused by a Skia change (https://codereview.chromium.org/1947153002/) that added a bunch of static initialization, which caused failures on these bots, e.g.:

https://build.chromium.org/p/chromium/builders/Mac/builds/15151
https://build.chromium.org/p/chromium/builders/Linux%20x64/builds/19052

I've created a revert to fix this: https://codereview.chromium.org/1948233002/

It reverts both the CL that introduced the problem (https://codereview.chromium.org/1947153002/) as well as a later CL that depends on that first CL (https://codereview.chromium.org/1948193002/).

reed@ has already reverted https://codereview.chromium.org/1948193002/, so I've instead created a separate revert of https://codereview.chromium.org/1947153002/ (revert is https://codereview.chromium.org/1945343003/).

The following revision refers to this bug:
  https://skia.googlesource.com/skia.git/+/d9f5d20f81b15190fa858eda1f536deae21e8e78

commit d9f5d20f81b15190fa858eda1f536deae21e8e78
Author: sclittle <sclittle@chromium.org>
Date: Thu May 05 01:23:30 2016

Revert of Modernize SkMutex and SkSemaphore. (patchset #2 id:20001 of https://codereview.chromium.org/1947153002/ )

Reason for revert:
This CL seems to have broken the Linux x64 and Mac
bots, e.g.

https://build.chromium.org/p/chromium/builders/Linux%20x64/builds/19052
https://build.chromium.org/p/chromium/builders/Mac/builds/15151

The error appears to have something to do with new static initializers being
added.

Original issue's description:
> Modernize SkMutex and SkSemaphore.
>
>    - use <atomic>
>    - fuse SkMutex and SkBaseMutex
>    - fuse SkSemaphore and SkBaseSemaphore
>
> Still TODO:
>    - replace SK_DECLARE_STATIC_MUTEX(name) with static SkMutex name
>
> I just didn't want to bother fixing all that up until I know this CL sticks.
>
> BUG=skia:
> GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1947153002
>
> No public API changes.
> TBR=reed@google.com
>
> Committed: https://skia.googlesource.com/skia/+/427c2819d9237d7d7729c59238036cfc73c072ea

TBR=herb@google.com,mtklein@chromium.org,reed@google.com,bsalomon@google.com
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG= 609340 

Review-Url: https://codereview.chromium.org/1945343003

[modify] https://crrev.com/d9f5d20f81b15190fa858eda1f536deae21e8e78/include/core/SkPixelRef.h
[modify] https://crrev.com/d9f5d20f81b15190fa858eda1f536deae21e8e78/include/ports/SkFontConfigInterface.h
[modify] https://crrev.com/d9f5d20f81b15190fa858eda1f536deae21e8e78/include/private/SkMutex.h
[modify] https://crrev.com/d9f5d20f81b15190fa858eda1f536deae21e8e78/include/private/SkSemaphore.h
[modify] https://crrev.com/d9f5d20f81b15190fa858eda1f536deae21e8e78/src/core/SkSemaphore.cpp
[modify] https://crrev.com/d9f5d20f81b15190fa858eda1f536deae21e8e78/src/lazy/SkDiscardableMemoryPool.cpp
[modify] https://crrev.com/d9f5d20f81b15190fa858eda1f536deae21e8e78/src/lazy/SkDiscardableMemoryPool.h

The revert in #4 should hopefully fix the "sizes" failure.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/423b16914aeb4988b7e4f1f64af6d77aa4d39086

commit 423b16914aeb4988b7e4f1f64af6d77aa4d39086
Author: skia-deps-roller <skia-deps-roller@chromium.org>
Date: Thu May 05 05:02:00 2016

Roll src/third_party/skia/ 7eb33da7e..d9f5d20f8 (1 commit).

https://chromium.googlesource.com/skia.git/+log/7eb33da7eede..d9f5d20f81b1

$ git log 7eb33da7e..d9f5d20f8 --date=short --no-merges --format='%ad %ae %s'
2016-05-04 sclittle Revert of Modernize SkMutex and SkSemaphore. (patchset #2 id:20001 of https://codereview.chromium.org/1947153002/ )

BUG= 609340 

CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
TBR=jvanverth@google.com

Review-Url: https://codereview.chromium.org/1948273003
Cr-Commit-Position: refs/heads/master@{#391759}

[modify] https://crrev.com/423b16914aeb4988b7e4f1f64af6d77aa4d39086/DEPS

This failure disappeared once the skia roll in #7 landed.

FYI, a build (c62a37ec0e514662c1418278b8a9a2707e38f336) for the second try of this is currently pending on both those builders.   I couldn't find trybots for them, but hope this time they'll pass.