Adds SHA256 Cryptocell code to bootloader

[kaczmarczyck]

The last piece to make the bootloader actually work. Advantage of a hardware SHA256 is the much faster boot time.

[jmichelp]

As we may need the Cryptocell code also in Tock, I wonder if this wouldn't make sense to make it a separate crate that we could import in both the bootloader and the kernel. Otherwise we'll end up with code duplication and different kinds of API to compute hashes.

[kaczmarczyck]

Question is if you want the new crate now or submit and put the work in when / if we submit it to Tock?

This PR is simpler than what I think you had in mind for Tock, i.e. it's API doesn't allow concurrent calls. So I don't know if a simply move to a new crate would actually be that useful?

[coveralls]

Coverage remained the same at 90.671% when pulling 5e30a47 on kaczmarczyck:bootloader into 91ba2c3 on google:develop.

[jmichelp]

We can merge it as-is at the moment to get the feature out.
And let's align with @ia0 to make it a separate crate that can bootstrap the work around hardware crypto support.

[ia0]

And let's align with @ia0 to make it a separate crate that can bootstrap the work around hardware crypto support.

Ack! (I don't have any ideas yet, and that might be related to ctap as a library too, so should definitely be in our radar).

[kaczmarczyck]

Thanks!