What do you mean by 'platform' here?

AFE correctly identifies the various models now:

Does mosys give the right model for these? 

pprabhu@pprabhu:files$ atest host list -b reef | ./contrib/count_labels -m
      7 basking
     38 electro
     17 reef

Do you mean goldeneye doesn't have the results under the right platforms? That will be the case for older builds where mosys wasn't reporting the correct model for us. But once mosys starts reporting the correct model, results in goldeneye from that build onwards should start showing up in the correct platform. Link to any recent results that are wrong?

Ignore the "does mosys give the right model for these?"

Leaked in from my initial response. mosys does, otherwise model: labels would have been wrong.

the special platform label that ends up in the tko database

I'll still need a link to something that is broken. Because aiui, the tko bit should also just work at this point for reef.

autotest doesn't have restful pages

but search for model:electro and you'll see that platform is set to 'reef' for all of them and we want it to be electro.

same for basking

The platform in AFE is unrelated to platform as understood by TKO (and hence goldeneye).

AFE's platform is (almost) dead. I don't know if there are some leftover users of it. We now use the model to populate 'platform' in the TKO results.
So, this will just work.

Please re-assign if you find a wmatrix / goldeneye view that has the wrong platform for either of these models. I still think this should just work, given that this is no different from coral.

This is dshi@'s CL that already handled this for coral, and I don't see anything new needed for reef: https://chromium-review.googlesource.com/767109

Example:

Test: http://cautotest/afe/#tab_id=view_job&object_id=159569136
Look at the .parse.log here: https://pantheon.corp.google.com/storage/browser/chromeos-autotest-results/159569136-chromeos-test/chromeos6-row3-rack10-host17

I do not completely understand how downstream result views us this (hence I asked dshi@ to write the CL), but looks like the 'platform' bit is still wrong (reef): http://shortn/_UjPy6Ic2y4

... but presumably this is not a problem for goldeneye and wmatrix (otherwise we'd have noticed in coral as well)?

+dshi

----------
Either I'm not understanding the request here, or the ask is premature.