Code review comment for lp:~smoser/maas-images/trunk.v3-formatting

I'd also like to raise an issue on the data being created here.

You are producing a kernel with "primary subarch" of
  hwe-16.04
and
  hwe-16.04-lowlatency

But you set their "subarches" to the same string.

How should one determine that one provides "low latency"?

When there is a hwe-16.10-lowlatency kernel, how will someone know that that would suffice if a user needed low latency ?

I'm not insisting that 'subarches' be used to indicate lowlatency, but i'm interested in how maas is knowing about this information.

It seems that we should add some metadata that explicitly sets the 'lowlatency' or generically "flavor" of the kernel.