flihp
commented
5 years ago High level feedback on this is very positive. This is the direction I wanted to take things and I really appreciate you pushing it further in this direction. A few questions that came to mind while reviewing this:
- Your commit message makes no mention of the sweeping changes you've made to the CI build in this work. If the CI changes are required to run the example applications under QEMU the bulk of these changes should be in a separate commit.
- Is it necessary for us to switch to docker and a different image (xenial -> debian buster)?
- There are a few bits of the environment that have been changed in the conversion to docker. Things like the choice for make
--jobsoption if changed need to be justified / explained.
Eventually I'd like to turn this into a sort of integration test harness much like we do in the tpm2-tss & tpm2-abrmd using the autotools mechanisms. Been thinking about it a bit and I do not think that turning this PR into that work is the right thing to do. Instead I'd like to merge this once we sort out separating the bulk of the changes to the CI from the introduction of the qemu execution in the CI (the whole startup.nsh thing is really nice, had no idea you could do that).
Run tpm2-get-caps-fixed.efi in QEMU after running build and test in travis