Chaining set_* methods in registers

[romancardenas]

Hi everyone!

First of all, I thought that it would be a nice idea to use Discussions for this topic. If you prefer an issue, let me know and I'll set up one.

@jsgf is working on a PR for riscv that adds a few set_* methods for read values of the mstatus register. I suggested following the cortex-m approach, in which these functions look like fn set_*(&mut self, val: T). However, he proposed using a fn set_*(&mut self, val: T) -> &mut Self fashion, so users could chain setter methods when configuring a register with multiple fields. I think this is a nice approach, but wanted to share it with the rest of the WG in case anyone has any objection (or if the cortex-m team would like to adopt this proposal too).