Some possible proposals: 

* Allow inheritance between overlays, such that the child model config overrides the parent.
* Define all models in the main "fizz" overlay and simply override kernel/software in child overlays and firmware in fizz overlay.

Or some other solution?

I'll take a look at this. Thanks for filing.

There's a few ways to do this.

Background
==========

The only fragment we care about sharing for Fizz is a few lines right here: https://chrome-internal.git.corp.google.com/chromeos/overlays/overlay-fizz-private/+/8fc884974fac790436d7ef28f02d8040493dd521/chromeos-base/chromeos-config-bsp/files/common/model.dtsi#33

The final binary configuration for an image is compiled by install_private_model_files here: https://chrome-internal.git.corp.google.com/chromeos/overlays/overlay-fizz-private/+/8fc884974fac790436d7ef28f02d8040493dd521/chromeos-base/chromeos-config-bsp/chromeos-config-bsp-9999.ebuild#31 . It walks all the .dtsi files installed to the build root, cats them together, and then compiles the final binary configuration format.

Short term
==========

Both of the short term approaches have the goal of only grabbing the common/model.dtsi file from a base.

1. Common baseboard: have a baseboard-fizz overlay that has just common/model.dtsi installed by chromeos-config-bsp-common (a new ebuild). overlay-fizz and overlay-fizz-moblab's chromeos-config-bsp ebuilds would depend on chromeos-config-bsp-common. During install_private_model_files, it finds the common/model.dtsi; sharing complete.

2. Inherited overlay: A new overlay would be created, overlay-fizz-moblab, that inherits from overlay-fizz. It would have a new ebuild, chromeos-config-bsp-moblab that installs moblab/accelerator specific configuration and depends on chromeos-config-bsp from the parent, overlay-fizz. chromeos-config-bsp in overlay-fizz would have a new USE flag "moblab" that enabled installing only the common/model.dtsi and skipping running install_private_model_files. install_private_model_files would run in chromeos-config-bsp-moblab thereby achieving sharing.

Neither of these would require support or new features from unibuild; they can be done immediately with no special knowledge or development work.

Long term
=========

3. We could add first-class support to the eclass for a special notion of shared configuration. This would be convention-based (i.e. files have special names or locations to indicate that they are shared).

4. We can use the new YAML-based configuration to pull in only sections of the YAML tree that we are interested in in the overridden/inheriting overlay.

Both of these would require some work and design and would take a month or so. We need to also account for private overlays that can't always be seen which adds additional complexity.

Let us know if you intend to skip options 1/2 and want us to go straight to 3/4. It would need to be prioritized against other things going on right now.

Option 1/2 look good enough in theory for Fizz/Wukong/Wukong-Accelerator/Wukong-Moblab

One other dimension to consider in the inheritance approach is Whitelabel.

Fizz/Wukong is WL, with a set of LOEMs/RLZs/Keys etc.
Fizz-Moblab is not WL, but it could be WL if that makes life easier, but with just one LOEM (Google)
Fizz-Accelerator is WL, with a different set of LOEMs/RLZs/Keys etc.

Does 1. and 2. 'just work' in this context?

Sorry for the delay. Yes, 1/2 just works in this context. Different builders means different keysets.

I've volunteered to implement Option 1 as soon as the overlays are created. Just let me know as soon as they are in place, and I'll implement a proof of concept.

I think they are all done,

chromeos/overlays/overlay-fizz-accelerator-private

although can't see fizz-moblab, although i know its being built.

Keith, where's fizz-moblab? i see guado-moblab but not fizz?

chromeos/overlays/overlay-variant-guado-moblab-private

Fizz moblab is in src/overlays/overlay-fizz-moblab/

I am not sure we have a private overlay,

I don't understand how moblab can be working at all at the moment. overlay-fizz-moblab inherits from baseboard-fizz: https://cs.corp.google.com/chromeos_public/src/overlays/overlay-fizz-moblab/profiles/base/parent which means that that build is completely eliding the unibuild architecture that fizz is using because overlay-fizz-private contains the unibuild configuration.

We definitely need the private overlay established and it should inherit from baseboard-fizz-private.

It looks like that would also pick up 1 other thing (redacted due to public bug) that moblab needs that is only in baseboard-fizz-private.

Who can set that up?

If I change overlay-fizz-moblab to have the parent baseboard-fizz-private will that solve the problem ?  or must we have a overlay-fizz-moblab-private ?

If your intention is that external contributors can build overlay-fizz-moblab, then they won't be able to do that if it inherits from an overlay that they can't see. Perhaps we don't need overlay-fizz-moblab at all?

Going back to #7/#8, does fizz-accelerator build look ok?

We will soon have a different FW version to use just for fizz-accelerator which should be a good test of the inheritance model?

Not at my desk right now but it looks like fizz-accelerator isn't in public or private Code Search?

per #8 i think its here,

chromeos/overlays/overlay-fizz-accelerator-private

I poked at my local checkout and confirmed that--at a glance--it looks right. I'm really busy this week so I might not get to this until later this week or first thing next week.

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chromeos/overlays/overlay-fizz-moblab-private/+/057507cd6f79385a80af439211f7fe0fa2c71d86

commit 057507cd6f79385a80af439211f7fe0fa2c71d86
Author: Keith Haddow <haddowk@chromium.org>
Date: Fri Feb 23 22:16:10 2018

Starting on this now.

for #10 - overlay-fizz-moblab-private is now set up and seems to be working correctly.

CL's converting overlay-fizz-private to a shared configuration are out for review.

Will send out -accelerator-private and -moblab-private conversion to unibuild with the shared fragment first thing in the morning.

moblab and accelerator are converted with shared configuration. All the CL's are out for review:

https://chromium-review.googlesource.com/c/chromiumos/overlays/chromiumos-overlay/+/940810
https://chrome-internal-review.googlesource.com/c/chromeos/overlays/baseboard-fizz-private/+/578550/2
https://chrome-internal-review.googlesource.com/c/chromeos/overlays/overlay-fizz-private/+/578551/2
https://chrome-internal-review.googlesource.com/c/chromeos/overlays/overlay-fizz-accelerator-private/+/579394
https://chrome-internal-review.googlesource.com/c/chromeos/overlays/overlay-fizz-moblab-private/+/579412

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/2440566e489bee168092708801ff014558972268

commit 2440566e489bee168092708801ff014558972268
Author: Jason D. Clinton <jclinton@chromium.org>
Date: Thu Mar 01 05:30:19 2018

eclass/cros-unibuild.eclass: Add support for shared baseboard config

There's a need for a fragment of configuration to be shared among
different overlays derived from the same baseboard. Implement this by
allowing a single file installed to
/usr/share/models/baseboard/model.dtsi to be added to the list of
sources to be compiled in to the final configuration.

BUG= chromium:801431 
TEST=emerge-fizz chromeos-config-bsp-baseboard chromeos-config-bsp
chromeos-config with the subsequent CL's

Change-Id: I7c54432a672ef55def8f889b1903a2adfa9a72b2
Reviewed-on: https://chromium-review.googlesource.com/940810
Commit-Ready: Jason Clinton <jclinton@chromium.org>
Tested-by: Jason Clinton <jclinton@chromium.org>
Reviewed-by: C Shapiro <shapiroc@google.com>

[modify] https://crrev.com/2440566e489bee168092708801ff014558972268/eclass/cros-unibuild.eclass

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chromeos/overlays/overlay-fizz-accelerator-private/+/03575ed86d20f92bac4abdb42ec68c8518a705a6

commit 03575ed86d20f92bac4abdb42ec68c8518a705a6
Author: Jason D. Clinton <jclinton@chromium.org>
Date: Fri Mar 02 00:13:16 2018

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chromeos/overlays/overlay-fizz-private/+/c068bebf973691f5328901d964b283daee70a961

commit c068bebf973691f5328901d964b283daee70a961
Author: Jason D. Clinton <jclinton@chromium.org>
Date: Fri Mar 02 00:13:22 2018

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chromeos/overlays/overlay-fizz-moblab-private/+/26d71274f754c516e15bed7bc9e72a111150ab39

commit 26d71274f754c516e15bed7bc9e72a111150ab39
Author: Jason D. Clinton <jclinton@chromium.org>
Date: Fri Mar 02 00:13:23 2018

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chromeos/overlays/baseboard-fizz-private/+/ddf615a207f8bea6739b911c4669e60d6fb2102e

commit ddf615a207f8bea6739b911c4669e60d6fb2102e
Author: Jason D. Clinton <jclinton@chromium.org>
Date: Fri Mar 02 00:13:15 2018

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chromeos/overlays/baseboard-fizz-private/+/a86ebb2d6423fb00668c02e659e8b40b44ccfb11

commit a86ebb2d6423fb00668c02e659e8b40b44ccfb11
Author: Jason D. Clinton <jclinton@chromium.org>
Date: Fri Mar 09 22:30:25 2018

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chromeos/overlays/overlay-fizz-private/+/d3c2d54a6daa5bf2316272612d73ad1f575c9579

commit d3c2d54a6daa5bf2316272612d73ad1f575c9579
Author: Jason D. Clinton <jclinton@chromium.org>
Date: Fri Mar 09 22:30:29 2018

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/ab5fc06491ea0a34b7397e7d586ef981014c3c09

commit ab5fc06491ea0a34b7397e7d586ef981014c3c09
Author: Jason D. Clinton <jclinton@chromium.org>
Date: Sun Mar 11 03:36:09 2018

eclass/cros-unibuild.eclass: Add support for shared baseboard config

There's a need for a fragment of configuration to be shared among
different overlays derived from the same baseboard. Implement this by
allowing a single file installed to
/usr/share/models/baseboard/model.dtsi to be added to the list of
sources to be compiled in to the final configuration.

BUG= chromium:801431 
TEST=emerge-fizz chromeos-config-bsp-baseboard chromeos-config-bsp
chromeos-config with the subsequent CL's

Change-Id: I7c54432a672ef55def8f889b1903a2adfa9a72b2
Reviewed-on: https://chromium-review.googlesource.com/940810
Commit-Ready: Jason Clinton <jclinton@chromium.org>
Tested-by: Jason Clinton <jclinton@chromium.org>
Reviewed-by: C Shapiro <shapiroc@google.com>
(cherry picked from commit 2440566e489bee168092708801ff014558972268)
Reviewed-on: https://chromium-review.googlesource.com/956922
Reviewed-by: Jason Clinton <jclinton@chromium.org>
Reviewed-by: YH Lin <yueherngl@chromium.org>
Commit-Queue: YH Lin <yueherngl@chromium.org>
Tested-by: YH Lin <yueherngl@chromium.org>

[modify] https://crrev.com/ab5fc06491ea0a34b7397e7d586ef981014c3c09/eclass/cros-unibuild.eclass

This merge is suspected of breaking the build on the R65 branch.

https://uberchromegw.corp.google.com/i/chromeos_release/builders/reef-android-nyc-pre-flight-branch%20release-R65-10323.B/builds/229

...
=== Start output for job chromeos-config-bsp-0.0.1-r1420 (0m5.6s) ===
chromeos-config-bsp-0.0.1-r1420: >>> Emerging (1 of 1) chromeos-base/chromeos-config-bsp-0.0.1-r1420::reef-private for /build/reef/
chromeos-config-bsp-0.0.1-r1420:  * Running stacked hooks for pre_pkg_setup
chromeos-config-bsp-0.0.1-r1420:  *    sysroot_build_bin_dir ...
chromeos-config-bsp-0.0.1-r1420:  [ ok ]
chromeos-config-bsp-0.0.1-r1420:  * Running stacked hooks for post_pkg_setup
chromeos-config-bsp-0.0.1-r1420:  *    python_eclass_hack ...
chromeos-config-bsp-0.0.1-r1420:  [ ok ]
chromeos-config-bsp-0.0.1-r1420:  * Running stacked hooks for pre_src_unpack
chromeos-config-bsp-0.0.1-r1420:  *    python_multilib_setup ...
chromeos-config-bsp-0.0.1-r1420:  [ ok ]
chromeos-config-bsp-0.0.1-r1420: >>> Unpacking source...
chromeos-config-bsp-0.0.1-r1420: Cloning into '/build/reef/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r1420/work/chromeos-config-bsp-0.0.1'...
chromeos-config-bsp-0.0.1-r1420: done.
chromeos-config-bsp-0.0.1-r1420: >>> Source unpacked in /build/reef/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r1420/work
chromeos-config-bsp-0.0.1-r1420:  * Running stacked hooks for post_src_unpack
chromeos-config-bsp-0.0.1-r1420:  *    asan_init ...
chromeos-config-bsp-0.0.1-r1420:  [ ok ]
chromeos-config-bsp-0.0.1-r1420: >>> Preparing source in /build/reef/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r1420/work/chromeos-config-bsp-0.0.1 ...
chromeos-config-bsp-0.0.1-r1420: >>> Source prepared.
chromeos-config-bsp-0.0.1-r1420: >>> Configuring source in /build/reef/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r1420/work/chromeos-config-bsp-0.0.1 ...
chromeos-config-bsp-0.0.1-r1420: >>> Source configured.
chromeos-config-bsp-0.0.1-r1420: >>> Compiling source in /build/reef/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r1420/work/chromeos-config-bsp-0.0.1 ...
chromeos-config-bsp-0.0.1-r1420: >>> Source compiled.
chromeos-config-bsp-0.0.1-r1420: >>> Test phase [not enabled]: chromeos-base/chromeos-config-bsp-0.0.1-r1420
chromeos-config-bsp-0.0.1-r1420: 
chromeos-config-bsp-0.0.1-r1420: >>> Install chromeos-config-bsp-0.0.1-r1420 into /build/reef/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r1420/image/ category chromeos-base
chromeos-config-bsp-0.0.1-r1420:  * Validating 0 files:
chromeos-config-bsp-0.0.1-r1420: usage: validate_config [-h] [-d] [-p] [-r] config [config ...]
chromeos-config-bsp-0.0.1-r1420: validate_config: error: too few arguments
chromeos-config-bsp-0.0.1-r1420:  * ERROR: chromeos-base/chromeos-config-bsp-0.0.1-r1420::reef-private failed (install phase):
chromeos-config-bsp-0.0.1-r1420:  *   Validation failed
chromeos-config-bsp-0.0.1-r1420:  * 
chromeos-config-bsp-0.0.1-r1420:  * Call stack:
chromeos-config-bsp-0.0.1-r1420:  *     ebuild.sh, line   93:  Called src_install
chromeos-config-bsp-0.0.1-r1420:  *   environment, line 3482:  Called install_private_model_files
chromeos-config-bsp-0.0.1-r1420:  *   environment, line 2535:  Called _install_model_files 'private-'
chromeos-config-bsp-0.0.1-r1420:  *   environment, line  416:  Called die
chromeos-config-bsp-0.0.1-r1420:  * The specific snippet of code:
chromeos-config-bsp-0.0.1-r1420:  *       validate_config -p "${files[@]}" || die "Validation failed";
chromeos-config-bsp-0.0.1-r1420:  * 
chromeos-config-bsp-0.0.1-r1420:  * If you need support, post the output of `emerge --info '=chromeos-base/chromeos-config-bsp-0.0.1-r1420::reef-private'`,
chromeos-config-bsp-0.0.1-r1420:  * the complete build log and the output of `emerge -pqv '=chromeos-base/chromeos-config-bsp-0.0.1-r1420::reef-private'`.
chromeos-config-bsp-0.0.1-r1420:  * The complete build log is located at '/build/reef/tmp/portage/logs/chromeos-base:chromeos-config-bsp-0.0.1-r1420:20180311-044358.log'.
chromeos-config-bsp-0.0.1-r1420:  * For convenience, a symlink to the build log is located at '/build/reef/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r1420/temp/build.log'.
chromeos-config-bsp-0.0.1-r1420:  * The ebuild environment file is located at '/build/reef/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r1420/temp/environment'.
chromeos-config-bsp-0.0.1-r1420:  * Working directory: '/build/reef/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r1420/work/chromeos-config-bsp-0.0.1'
chromeos-config-bsp-0.0.1-r1420:  * S: '/build/reef/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r1420/work/chromeos-config-bsp-0.0.1'
chromeos-config-bsp-0.0.1-r1420: >>> Failed to emerge chromeos-base/chromeos-config-bsp-0.0.1-r1420 for /build/reef/, Log file:
chromeos-config-bsp-0.0.1-r1420: >>>  '/build/reef/tmp/portage/logs/chromeos-base:chromeos-config-bsp-0.0.1-r1420:20180311-044358.log'
chromeos-config-bsp-0.0.1-r1420: 
chromeos-config-bsp-0.0.1-r1420:  * Messages for package chromeos-base/chromeos-config-bsp-0.0.1-r1420 merged to /build/reef/:
chromeos-config-bsp-0.0.1-r1420: 
chromeos-config-bsp-0.0.1-r1420:  * ERROR: chromeos-base/chromeos-config-bsp-0.0.1-r1420::reef-private failed (install phase):
chromeos-config-bsp-0.0.1-r1420:  *   Validation failed
chromeos-config-bsp-0.0.1-r1420:  * 
chromeos-config-bsp-0.0.1-r1420:  * Call stack:
chromeos-config-bsp-0.0.1-r1420:  *     ebuild.sh, line   93:  Called src_install
chromeos-config-bsp-0.0.1-r1420:  *   environment, line 3482:  Called install_private_model_files
chromeos-config-bsp-0.0.1-r1420:  *   environment, line 2535:  Called _install_model_files 'private-'
chromeos-config-bsp-0.0.1-r1420:  *   environment, line  416:  Called die
chromeos-config-bsp-0.0.1-r1420:  * The specific snippet of code:
chromeos-config-bsp-0.0.1-r1420:  *       validate_config -p "${files[@]}" || die "Validation failed";
chromeos-config-bsp-0.0.1-r1420:  * 
chromeos-config-bsp-0.0.1-r1420:  * If you need support, post the output of `emerge --info '=chromeos-base/chromeos-config-bsp-0.0.1-r1420::reef-private'`,
chromeos-config-bsp-0.0.1-r1420:  * the complete build log and the output of `emerge -pqv '=chromeos-base/chromeos-config-bsp-0.0.1-r1420::reef-private'`.
chromeos-config-bsp-0.0.1-r1420:  * The complete build log is located at '/build/reef/tmp/portage/logs/chromeos-base:chromeos-config-bsp-0.0.1-r1420:20180311-044358.log'.
chromeos-config-bsp-0.0.1-r1420:  * For convenience, a symlink to the build log is located at '/build/reef/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r1420/temp/build.log'.
chromeos-config-bsp-0.0.1-r1420:  * The ebuild environment file is located at '/build/reef/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r1420/temp/environment'.
chromeos-config-bsp-0.0.1-r1420:  * Working directory: '/build/reef/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r1420/work/chromeos-config-bsp-0.0.1'
chromeos-config-bsp-0.0.1-r1420:  * S: '/build/reef/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r1420/work/chromeos-config-bsp-0.0.1'
=== Complete: job chromeos-config-bsp-0.0.1-r1420 (0m5.6s) ===
...

Timing wise https://chromium-review.googlesource.com/#/c/chromiumos/overlays/chromiumos-overlay/+/956922/ landed saturday evening, and it fits in the timeline for the first failed build, and has vague connections to the failure.

This bug also does not appear to have been merge approved.

The CL will be reverted.

FWIW 66 appears to be broken also, and has been since the branch point, not sure if we want to just revert there too, if 65 comes back green I would advise doing so.

Yes, we should revert once M65 revert is proven to work 

jclinton@, Going forward please make sure to use merge-request labels and wait for approval  prior to any merge to the branches.
Once you have a working CL we can evaluate merge again. 

I think this was accidentally cherry-picked back by someone else as part of a firmware uprev. It probably wasn't intentional. I certainly didn't have any desire to backport it and I don't think anyone else did either. Not sure it makes sense to backport unless it somehow affected -accelerator FSI.

Sorry, I did the cherry-pick in the first place. My apologies!!

My local build completed without problem, and since there're few changes (yes, related to FW rev) sitting on of it I tried to bring them in sync. Guess I was wrong...

I think there is other stuff that is broken by this, Fizz has not built for a few days now on 65.

...
chromeos-config-bsp-0.0.1-r28: Failed: return code: 2; command: dtc -I dts -o /build/fizz/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r28/temp/tmpJDR_QE.dtb -O dtb /build/fizz/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r28/temp/tmpXeppXg.dts
chromeos-config-bsp-0.0.1-r28: /build/fizz/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r28/temp/tmpJDR_QE.dtb: ERROR (phandle_references): Reference to non-existent node or label "shared"
chromeos-config-bsp-0.0.1-r28: 
chromeos-config-bsp-0.0.1-r28: /build/fizz/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r28/temp/tmpJDR_QE.dtb: ERROR (phandle_references): Reference to non-existent node or label "shared"
chromeos-config-bsp-0.0.1-r28: 
chromeos-config-bsp-0.0.1-r28: /build/fizz/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r28/temp/tmpJDR_QE.dtb: ERROR (phandle_references): Reference to non-existent node or label "shared"
chromeos-config-bsp-0.0.1-r28: 
chromeos-config-bsp-0.0.1-r28: /build/fizz/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r28/temp/tmpJDR_QE.dtb: ERROR (phandle_references): Reference to non-existent node or label "shared"
chromeos-config-bsp-0.0.1-r28: 
chromeos-config-bsp-0.0.1-r28: /build/fizz/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r28/temp/tmpJDR_QE.dtb: ERROR (phandle_references): Reference to non-existent node or label "shared"
chromeos-config-bsp-0.0.1-r28: 
chromeos-config-bsp-0.0.1-r28: ERROR: Input tree has errors, aborting (use -f to force output)
chromeos-config-bsp-0.0.1-r28: 
chromeos-config-bsp-0.0.1-r28: cmd=['dtc', '-I', 'dts', '-o', '/build/fizz/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r28/temp/tmpJDR_QE.dtb', '-O', 'dtb', '/build/fizz/tmp/portage/chromeos-base/chromeos-config-bsp-0.0.1-r28/temp/tmpXeppXg.dts']
chromeos-config-bsp-0.0.1-r28:  * ERROR: chromeos-base/chromeos-config-bsp-0.0.1-r28::fizz-private failed (install phase):
chromeos-config-bsp-0.0.1-r28:  *   Validation failed
...

https://crosland.corp.google.com/log/10323.54.0..10323.55.0

If we want Fizz to have a build on 65, we need to fix this ASAP, Fizz has officially missed the first stable release.

Should we revert the entire set of https://chrome-internal-review.googlesource.com/#/q/topic:fizz-r65-cherry-picks+(status:open+OR+status:merged) ?

We can revert the entire set but the signing and the changes (key_id mapping) will be needed.

Let me see if there's anything I can do.

Thanks.

I have started a tryjob to see if this works better.

https://uberchromegw.corp.google.com/i/chromiumos.tryserver/builders/release/builds/21048

will update once it's done.

Thanks.

Found a mistake, aborted and restarted the tryjob,

https://uberchromegw.corp.google.com/i/chromiumos.tryserver/builders/release/builds/21050

Thanks.

https://uberchromegw.corp.google.com/i/chromiumos.tryserver/builders/release/builds/21050

tryjob completes successfully.

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chromeos/overlays/overlay-fizz-private/+/3f0830235434ebc7466942e770da8fb6de3c705a

commit 3f0830235434ebc7466942e770da8fb6de3c705a
Author: YH Lin <yueherngl@google.com>
Date: Fri Mar 16 21:36:22 2018

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chromeos/overlays/overlay-fizz-private/+/c177d24af96838a052881504e5489997f1546bce

commit c177d24af96838a052881504e5489997f1546bce
Author: YH Lin <yueherngl@google.com>
Date: Tue Mar 20 17:44:05 2018

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chromeos/overlays/overlay-fizz-moblab-private/+/202ddd6a1f59089c461616cba6ada134113533d6

commit 202ddd6a1f59089c461616cba6ada134113533d6
Author: YH Lin <yueherngl@google.com>
Date: Wed Mar 21 17:26:09 2018

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chromeos/overlays/overlay-fizz-accelerator-private/+/3ba3ffd4a828a0dc5c7796ae8f38179e8b9b7b26

commit 3ba3ffd4a828a0dc5c7796ae8f38179e8b9b7b26
Author: YH Lin <yueherngl@google.com>
Date: Wed Mar 21 02:31:55 2018

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chromeos/overlays/baseboard-fizz-private/+/2496b1fe0892910bb3d23e945844cd4a94c3e70d

commit 2496b1fe0892910bb3d23e945844cd4a94c3e70d
Author: Jason D. Clinton <jclinton@chromium.org>
Date: Fri Mar 23 18:16:41 2018

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chromeos/overlays/baseboard-fizz-private/+/bf300e02480db033de96ab111659ced5e9cb718a

commit bf300e02480db033de96ab111659ced5e9cb718a
Author: Jason D. Clinton <jclinton@chromium.org>
Date: Fri Mar 23 18:20:09 2018

This was implemented quite a long time ago.