Is the change in 04_zero_overhead_abstraction valid? about rust-raspberrypi-os-tutorials HOT 3 CLOSED

This is a very interesting question, but also remember small functions may not use the stack.

This does come up - if I compile with opt-level=0 then I think my code hangs for this reason.

from rust-raspberrypi-os-tutorials.

You‘re right.

Setting the stack pointer should ideally be done in early boot assembly before the first function call happens.
You cannot rule out that the compiler does not honor the inline attributes in this code snippet.

I threw out the assembly here nevertheless, mostly “because we can”, because I like to have as much Rust code as possible in this toy-kernel 😎.
We have emulation for testing and a corresponding CI pipeline that would raise a flag if the compiler inserts function calls here, because the booting would fail early on.
Also, for functions that boil down to very few lines of assembly (in the SP case a single one), you can be quite optimistic that the compiler honors the inline suggestion.

In the original version of this particular tutorial I mentioned that we would need to manually check that function calls don’t happen here.
I wanted to mention it in the new version here too, it didn’t come around porting the text over yet.

Hope that helps!

Best,
Andre

from rust-raspberrypi-os-tutorials.

@andre-richter -- thx for the explanation and the pointer to the original post!

mostly “because we can”, because I like to have as much Rust code as possible in this toy-kernel sunglasses.

makes sense :)

from rust-raspberrypi-os-tutorials.