The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/ee8618f9e765917145cc20b136cb6a3ba324db53

commit ee8618f9e765917145cc20b136cb6a3ba324db53
Author: charliea <charliea@chromium.org>
Date: Fri Dec 09 15:43:55 2016

Make the BattOr agent retry start tracing when we fail to read an ack

This should greatly mitigate a current bug in the BattOr firmware where
sending the command to start tracing causes the BattOr to reset. In this
case, we now retry the full initialization sequence. Because the error
we're hitting is a timeout, and we allow 4 seconds for a command to
be executed before timing out, this means that initialization will
retry after 4 seconds and then make 20 attempts to initialize the BattOr
100 milliseconds apart from each other, making the whole sequence take
something like 6 additional seconds. We can retry this full sequence
up to 5 times. When added to the five or so seconds that the BattOr
already takes to initialize, this means that this full sequence might
take up to 11 * 5 = 55 seconds before deciding that the BattOr can't
be initialized. In practice, I suspect that almost all initializations
will succeed on either their first or second attempts.

BUG= 672631 

Review-Url: https://codereview.chromium.org/2563033002
Cr-Commit-Position: refs/heads/master@{#437548}

[modify] https://crrev.com/ee8618f9e765917145cc20b136cb6a3ba324db53/tools/battor_agent/battor_agent.cc
[modify] https://crrev.com/ee8618f9e765917145cc20b136cb6a3ba324db53/tools/battor_agent/battor_agent.h
[modify] https://crrev.com/ee8618f9e765917145cc20b136cb6a3ba324db53/tools/battor_agent/battor_agent_unittest.cc
[modify] https://crrev.com/ee8618f9e765917145cc20b136cb6a3ba324db53/tools/battor_agent/battor_error.cc
[modify] https://crrev.com/ee8618f9e765917145cc20b136cb6a3ba324db53/tools/battor_agent/battor_error.h

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/9f0a98a3c001f7602338311a4765eaac0bed4376

commit 9f0a98a3c001f7602338311a4765eaac0bed4376
Author: catapult-deps-roller <catapult-deps-roller@chromium.org>
Date: Sat Dec 10 04:02:47 2016

Roll src/third_party/catapult/ 707aaac64..19565fdb1 (2 commits).

https://chromium.googlesource.com/external/github.com/catapult-project/catapult.git/+log/707aaac64b3c..19565fdb148a

$ git log 707aaac64..19565fdb1 --date=short --no-merges --format='%ad %ae %s'
2016-12-09 nednguyen Revert of re-enabled netlog viewer dev server tests (patchset #2 id:20001 of https://codereview.chromium.org/2553213002/ )
2016-12-09 charliea Roll battor_agent_bin to include new init retry logic

BUG= 672631 

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, see:
http://www.chromium.org/developers/tree-sheriffs/sheriff-details-chromium#TOC-Failures-due-to-DEPS-rolls

CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.android:android_optional_gpu_tests_rel
TBR=catapult-sheriff@chromium.org

Review-Url: https://codereview.chromium.org/2567543003
Cr-Commit-Position: refs/heads/master@{#437746}

[modify] https://crrev.com/9f0a98a3c001f7602338311a4765eaac0bed4376/DEPS

Issue 650426 has been merged into this issue.

I'm seeing some battor failures on a windows bot. Could this be the same issue?

stdio link: https://luci-logdog.appspot.com/v/?s=chrome%2Fbb%2Fchromium.perf%2FWin_10_High-DPI_Perf%2F75%2F%2B%2Frecipes%2Fsteps%2Fbattor.power_cases_on_Intel_GPU_on_Windows_on_Windows-10-10240%2F0%2Fstdout

Yes, those appear to be related to this.

Marking this as fixed now that we have the init retry workaround logic in place.