See also: https://bugs.chromium.org/p/chromium/issues/detail?id=780516

Some variant:XXXX labels also seem to have the platform bit set.

 Issue 782507  has been merged into this issue.

The CL to fix model labels for old DUTs was pushed today. I'll check tomorrow if they were updated correctly. This is blocked on all (most) DUTs getting the correct model information.

Rinse and repeat: crbug.com/555675

I think my understanding here was wrong. wmatrix's (and possibly goldeneye's?) 'platform' is actually determined by tko entires only. And we seem to be dumping the platform label based on some logic in tko/parse: 

https://chromium-review.googlesource.com/c/chromiumos/third_party/autotest/+/312566

So... that's what I update to dump model if a model is available? (falling back to board if model is not available, for non-cros boards)

-->dshi@ to confirm.

TKO records are generated based on test keyvals. The board information is retrieved from board label which is included in job keyvals:
https://cs.corp.google.com/android/external/autotest/tko/parse.py?rcl=e6064014132c17d4893ee165c0836357f7258231&l=397

is there a path decided on here .... coral testing is blocked by this

pprabhu is working on this

Dan has agreed to do this early next week. It will be faster that way, seriously.

I realize that I don't know quite enough about TKO to do this in a hurry. I'll follow along for future requests of this sort.

Following CL is landed and pushed to lab, tko should be using model: label for machine_group now.
https://chromium-review.googlesource.com/#/c/chromiumos/third_party/autotest/+/767109/