RAM-based interrupt vector tables

[9names]

To improve performance for interrupts and allow for dynamically hooking interrupt vectors at runtime, provide some facilities for creating and using interrupt vector tables.
This should also increase what we can do with multicore.

This code is currently completely untested.

[jannic]

Didn't test the code either, but it looks sensible.
Maybe find a better place, though? I don't think it's multicore-related.

[9names]

I have tested it since I wrote the first comment, but not in a while. It explodes.
Haven't gotten around to working out why yet.

I don't think it's multicore-related.

You're right. I had written this to help out with a multicore experiment so that is what was on my mind.
I'll need to consider where it should go.

[9names]

I've rebased this and written an example testing it.
It seems to work fine now, despite my previous comment stating that it exploded.
Not sure if if needs a fence after the VTOR copy.

Also decided to go back to just copying the 192 bytes that are used by the table entries instead of core::mem::size_of::<VectorTable>().
I still don't have a good mental model of whether the padding bytes (256 bytes - 192) will be used by other static variables, so avoiding writing to them seems the safest option.

[jannic]

I still don't have a good mental model of whether the padding bytes (256 bytes - 192) will be used by other static variables, so avoiding writing to them seems the safest option.

This seems to be an open question:

• Separate size and stride for types rust-lang/rfcs#1397
• Remove size != stride distinction for array layout rust-lang/unsafe-code-guidelines#176

[9names]

Thanks for those threads. I think the answer (right now) is that the padding bytes are unused.
And, it doesn't seem like any existing target will support stride != size, so that should hold in the future.
Would be nice to have a way to overalign the start of an array to avoid this though - maybe we'll get magic syntax for that some day but I don't want it so badly that I'll push for it.